We briefly list all mobile graphics cards currently available. The GPUs are sorted by performance and then separated into different classes. To get a better overview on current graphics cards, older cards can be greyed out.
High-End Graphics Cards - These graphics cards are able to play the latest and most demanding games in high resolutions and full detail settings with enabled Anti-Aliasing.
A Blackwell family graphics card for laptops, that Team Green unveiled during CES 2025. We expect it to be based on the GB203 chip with 10,496 CUDA cores active.
Blackwell family graphics card for laptops that Team Green unveiled during CES 2025. We expect it to be based on the GB203 chip with 7680 unified shaders (CUDA cores) active.
High-end laptop graphics card based on the Ada Lovelace architecture. Offers 16 GB GDDR6 VRAM with a 256 Bit memory bus and can be configured with 80 - 150 TDP settings (+ Dynamic Boost).
A super-powerful professional graphics card for use in laptops that sports 9,728 CUDA cores and 16 GB of ECC GDDR6 VRAM. Brought into existence in 2023, this graphics adapter leverages TSMC's 5 nm process (4N) and Nvidia's Ada Lovelace architecture to achieve great performance combined with reasonable power consumption. Hardware-wise, the RTX 5000 Ada is a GeForce RTX 4090 Laptop in disguise; consequently, both make use of the AD103 chip and have little difficulty running triple-A games at UHD 2160p. The Nvidia-recommended TGP range is modestly wide at 80 W to 175 W leading to noticeable performance differences between different systems powered by what is supposed to be the same graphics card.
Fast high-end mobile graphics card based on the RDNA 3 architecture. Offers 4,608 cores (72 CUs) at a TDP of up to 200 Watt (including SmartShift). Uses the Navi 31 chip that is manufactured in 5nm at TSMC.
High-end graphics card for laptops based on the Ada Lovelace architecture. Offers 12 GB GDDR6 (192 Bit) and is configurable with a TGP of 60 - 150W (+ Dynamic Boost).
A very powerful professional graphics card for use in laptops that sports 7,424 CUDA cores and 12 GB of ECC GDDR6 VRAM. Brought into existence in 2023, this graphics adapter leverages TSMC's 5 nm process (4N) and Nvidia's Ada Lovelace architecture to achieve great performance combined with moderate power consumption. Hardware-wise, the RTX 4000 is a GeForce RTX 4080 Laptop in disguise; consequently, both make use of the AD103 chip and have little difficulty running triple-A games at QHD 1440p. The Nvidia-recommended TGP range is modestly wide at 60 W to 175 W leading to bizarre performance differences between different systems powered by what is supposed to be the same graphics card.
7424 - unified, DX12 Ultimate | 16000 MHz, 192 Bit
A Blackwell family graphics card for laptops that Team Green unveiled during CES 2025. We expect it to be based on either the GB204 or the GB205 chip, since, according to Nvidia, it contains 5888 active CUDA cores.
A higher-end professional graphics card for use in laptops that sports 5,120 CUDA cores and 12 GB of ECC GDDR6 VRAM. Brought into existence in 2023, this graphics adapter leverages TSMC's 5 nm (4N) process and Nvidia's Ada Lovelace architecture; hardware-wise, the RTX 3500 is a cut-down GeForce RTX 4070 (Desktop), as far as we can tell. Consequently, it makes use of the same AD104 chip and will have little difficulty running triple-A games at QHD 1440p. The Nvidia-recommended TGP range for the card is fairly wide at 60 W to 140 W leading to noticeable performance differences between different systems powered by what is supposed to be the same graphics card.
5120 - unified, DX12 Ultimate | 16000 MHz, 192 Bit
High-End laptop graphics card based on the Ampere GA103S chip. At the time of the launch in early 2022, the 3080 Ti is the fastest laptop graphics card on the market.
A pretty fast integrated graphics adapter designed by Apple that's built into the 16-core Apple M4 Max SoC. It features 40 "cores" and has access to at least 48 GB of fast 546 GB/s LPDDR5x on-package RAM depending on the configuration. Its gaming performance is set to be about as good as that of the RTX 4070 Laptop.
High-end graphics card for laptops based on the Ada Lovelace architecture. Offers 8 GB GDDR6 (128 Bit) and is configurable with a TGP of 35 - 115W (+ Dynamic Boost).
A higher-end professional graphics card for use in laptops that sports 4,608 CUDA cores and 8 GB of ECC GDDR6 VRAM. It would be fair to say that this is a GeForce RTX 4070 (Laptop) in disguise. Brought into existence in 2023, the RTX 3000 leverages TSMC's 5 nm process and Nvidia's Ada Lovelace architecture to achieve very decent performance combined with moderate power consumption. This graphics card is fast enough to run many games at 1440p with quality set to High. The Nvidia-recommended TGP range is very wide at 35 W to 140 W leading to bizarre performance differences between different systems powered by what is supposed to be the same graphics card.
4608 - unified, DX12 Ultimate | 16000 MHz, 128 Bit
High-End professional laptop graphics card based on the Ampere GA103S chip. Based on the consumer GeForce GTX 3080 Ti with certified drivers. Available in different performance variants ranging from 80 - 165 W TGP (max. power consumption).
High-End laptop graphics card based on the Ampere GA104 chip. At the time of the launch in early 2021, the 3080 is the fastest laptop graphics card on the market.
High-end mobile graphics card based on a Navi 22 chip (new RDNA2 architecture) and manufactured in the modern 7nm process. Offers 2,560 cores (40 CUs) and using 12 GB GDDR6 graphics memory. The game clock is specified at 2.463 GHz at 135-165 Watt TDP.
High-End laptop graphics card based on the Ampere GA104 chip. Available in different versions with a TDP from 80 to 125 Watt and different maximum Boost clock speeds of 1035 - 1485 MHz.
High-End professional laptop graphics card based on the Ampere GA104 chip. Based on the consumer GeForce GTX 3070 Ti with certified drivers. Available in different performance variants ranging from 80 - 140 W TGP (max. power consumption).
High-End professional laptop graphics card based on the Ampere GA104 chip. Based on the consumer GeForce GTX 3080 with certified drivers. Available in different performance variants ranging from 80 - 165 W TGP (max. power consumption).
High-end graphics card for laptops based on the Ada Lovelace architecture. Offers 8 GB GDDR6 (128 Bit) and is configurable with a TGP of 35 - 115W (+ Dynamic Boost).
A mid-range professional graphics card for use in laptops that sports 3,072 CUDA cores and 8 GB of GDDR6 VRAM. It would be fair to say that this is a GeForce RTX 4060 (Laptop) in disguise; consequently, the former is powered by the same AD107 chip as the latter, and is fast enough to handle any triple-A game at 1080p with Ultra quality settings. The product was brought into existence in 2023. It leverages TSMC's 5 nm process and the Ada Lovelace architecture; the Nvidia-recommended TGP range is very wide at 35 W to 140 W leading to bizarre performance differences between different systems powered by what is supposed to be the same graphics card.
3072 - unified, DX12 Ultimate | 16000 MHz, 128 Bit
Professional high-end laptop graphics card based on the TU102 chip with 4,608 shaders and 24 GB GDDR6 VRAM. Compared to the desktop variant, the GPU is clocked lower. Nvidia states that a desktop system with RTX 6000 is about 13% faster in SpecViewPerf. Due to the high power consumption of 200 Watt, the RTX 6000 is only suited for big laptops with very good cooling solutions.
High-end mobile graphics card based on a Navi 22 chip (new RDNA2 architecture) and manufactured in the modern 7nm process. Offers 2,560 cores (40 CUs) and using 12 GB GDDR6 graphics memory. The game clock is specified at 2.3 GHz at 145 Watt TDP.
High-End laptop graphics card based on the TU104 chip. At the time of the launch in early 2020, the 2080 Super is the fastest laptop graphics card on the market, but only slightly faster than the old GeForce RTX 2080.
High-End professional laptop graphics card based on the Ampere GA104 chip. Based on the consumer GeForce GTX 3070 with certified drivers. Available in different performance variants ranging from 80 - 140 W TGP (max. power consumption).
High-end mobile graphics card based on a Navi 23 chip (new RDNA2 architecture) and manufactured in the modern 7nm process. Offers 2,048 cores (32 CUs) and using 12 GB GDDR6 graphics memory. The game clock is specified at 2 GHz at 100 Watt TDP. The 6800S is the power saving variant of the Radeon RX 6800M.
An upcoming Blackwell family graphics card for laptops that Team Green is very likely to unveil during CES 2025. Considering the 4060 Laptop has 3,072 unified shaders (CUDA cores) at its disposal, the 5060 will most most likely sport over 4,000 shaders.
Fast mid-range mobile graphics card based on the RDNA 3 architecture. Offers 2,048 cores (32 CUs) at a TDP of up to 120 Watt. Uses the Navi 33 chip that is manufactured in 6nm at TSMC.
Fast mid-range mobile graphics card based on the RDNA 3 architecture. Offers 2,048 cores (32 CUs) at a TDP of up to 100 Watt. Uses the Navi 33 chip that is manufactured in 6nm at TSMC.
Integrated graphics card offering all 38 cores designed by Apple and integrated in the Apple M2 Max SoC. The graphics card has no dedicated graphics memory but can use the fast LPDDR5-6400 unified memory with a 512 bit bus (up to 400 GBit/s).
High-End laptop graphics card based on the TU104 chip with 2,944 shaders and 8 GB GDDR6 VRAM. Compared to the desktop variant, the GPU is clocked lower.
Mid-range mobile graphics card based on a Navi chip (new RDNA2 architecture most likely) and manufactured in the modern 7nm process. Offers 2,304 cores (36 CUs) and using GDDR6 graphics memory.
Professional high-end laptop graphics card based on the TU104 chip with 3,072 shaders and 16 GB GDDR6 VRAM. Compared to the desktop variant, the GPU is clocked lower. There are different versions available with 110 Watt (1035 - 1545 MHz) and 150 Watt (1350 - 1770 MHz) in addition to the even lower clocked Max-Q variant.
Fast mid-range mobile graphics card based on the RDNA 3 architecture. Offers 1,792 cores (28 CUs) at a TDP of up to 90 Watt. Uses the Navi 33 chip that is manufactured in 6nm at TSMC.
A pretty fast integrated graphics adapter designed by Apple that's built into the 14-core Apple M4 Max SoC. It features 32 "cores" and has access to at least 36 GB of fast 410 GB/s LPDDR5x on-package RAM depending on the configuration. Its gaming performance is set to be about as good as that of the RTX 4050 Laptop.
Integrated graphics card offering 32 cores (4096 ALUs) designed by Apple and integrated in the Apple M1 Max SoC. The theoretical performance is 10.4 Teraflops and it can use the 32 or 64 GB or unified memory.
Fast mid-range mobile graphics card based on the RDNA 3 architecture. Offers 1,792 cores (28 CUs) at a TDP of up to 90 Watt. Uses the Navi 33 chip that is manufactured in 6nm at TSMC. Low-power variant of the RX 7600M with slightly reduced performance.
High-End laptop graphics card based on the TU104 chip with 3,072 shaders. Runs at reduced clock speeds compared to a laptop RTX 2080 Super to achieve a much lower power consumption. The TGP is specified between 80 and 90 Watts with different clock speeds.
Mid-range professional laptop graphics card based on the Ampere GA104 chip. Similar to the consumer GeForce GTX 3060 with certified drivers. Available in different performance variants ranging from 60 - 130 W TGP (max. power consumption).
Integrated graphics card offering 24 of the 32 cores in the M1 Max chip. It offers no dedicated graphics memory but can use the fast unified memory (400 GB/s 512 Bit LPDDR5-6400).
The Intel Arc A770M is a dedicated mobile lower mid-range graphics card based on Xe HPG microarchitecture. It offers 32 cores (384 ALUs), 32 ray tracing units and a base core of 1650 MHz. The TGP is specified between 120 and 150 Watt.
Mid-range mobile graphics card based on the Navi architecture. Offers 2,048 cores (32 CUs) at a TDP of up to 120 Watt. Uses the Navi 23 chip based on the RDNA 2 architecture.
Graphics card designed by Apple itself in the Apple M2 Pro SoC and, according to Apple, with all 18 cores. The GPU uses a new architecture and now offers dynamic caching, mesh shading and ray tracing acceleration.
Mid-range mobile graphics card based on the Navi architecture. Offers 1792 cores (28 CUs) at a TDP of 80 - 120 Watt. Uses the Navi 23 chip based on the RDNA 2 architecture.
Mid-range mobile graphics card based on the Navi architecture. Offers 1792 cores (28 CUs) at a TDP of 90 - 135 Watt. Uses the Navi 23 chip based on the RDNA 2 architecture and can be configured by the laptop manufacturer from 50 - 100 Watt TDP.
Mid-range mobile graphics card based on a Navi chip (new RDNA2 architecture most likely) and manufactured in the modern 7nm process. Offers 1,792 cores (28 CUs) and using 8 GB GDDR6 graphics memory. The 6700S is the power saving variant of the 6700M.
High-End laptop graphics card based on the TU106 chip with 2,304 shaders and 8 GB GDDR6 VRAM. Compared to the desktop variant, the GPU is clocked lower. The refresh version in April 2020 bears the same name, but is slightly higher clocked.
Professional high-end laptop graphics card based on the TU104 chip with 3,072 shaders and 16 GB GDDR6 VRAM. Compared to the normal mobile variant, the Max-Q variants are clocked lower and work in a more efficient state. Currently, we know of three variants with different clock speeds and power consumptions (80, 85 and 90 W).
High-End laptop graphics card based on the TU104 chip with 2,944 shaders and 8 GB GDDR6 VRAM. Runs at reduced clock speeds compared to a laptop RTX 2080 to achieve a much lower power consumption.
The RX5700M is a mobile high-end graphics card for laptops. It uses the Navi 10 chip (like the desktop RX 5700 and 5700 XT) and therefore also the RDNA architecture. Like the desktop counterparts, the mobile RX 5700M uses dedicated GDDR6 graphics memory connected with a 256-bit bus. The performance is intended to be between the mobile RTX 2070 and RTX 2080 (3DMark 11 GPU score).
Pascal based laptop graphics card using GDDR5 graphics memory. Should offer a similar performance to the deskto GTX 1080. However, there is also a more power efficient "Max-Q" version that performs worse than the normal GTX1080 and is used in thin and light laptops.
Professional high-end laptop graphics card based on the TU104 chip with 2,560 shaders and 8 GB GDDR6 VRAM. Compared to the desktop variant, the GPU is clocked lower. Compared to consumer RTX cards, the Quadro RTX 4000 is positioned between the RTX 2070 and 2080.
High-End laptop graphics card based on the TU104 chip with 2,560 shaders. Runs at reduced clock speeds compared to a laptop RTX 2070 Super to achieve a much lower power consumption. The TGP is specified at 80 Watts and therefore significantly lower than the RTX 2070 Super Mobile (115 Watt).
Professional high-end laptop graphics card based on the TU104 chip with 2,560 shaders and 8 GB GDDR6 VRAM. Compared to the desktop variant, the GPU is clocked lower. Compared to the normal mobile variant, the Max-Q variants are clocked lower and work in a more efficient state. Currently we know of three variants with different clock speeds and power consumptions (80, 85 and 90 W).
High-End laptop graphics card based on the TU106 chip with 2,304 shaders and 8 GB GDDR6 VRAM. Offers reduced clock speeds to achieve a much lower power consumption.
Mid-range mobile graphics card based on the Navi architecture. Offers 1792 cores (28 CUs) at a TDP of up to 80 Watt. Uses the Navi 23 chip based on the RDNA 2 architecture and is the power saving variant of the RX 6600M.
High-end workstation graphics card for laptops that is based on the 16 nm GP104 chip (Pascal). Similar to the desktop GTX 1080 but with GDDR5 memory. An efficient but slower Max-Q variant is also available.
High-end professional laptop graphics card based on the consumer GeForce RTX 2070 (TU106 chip) of with 2,304 shaders and 8 GB GDDR6 VRAM. Compared to the desktop variant, the GPU is clocked lower. It offers certified drivers that are optimized for stability and performance.
An upcoming Blackwell family graphics card for laptops that Team Green is very likely to unveil during CES 2025. Considering the 4050 Laptop has 2,560 unified shaders (CUDA cores) at its disposal, the 5050 will most most likely sport 3,072 or 3,584 shaders.
Mid range to high end laptop graphics card based on the Turing architecture with raytracing features and 1,920 shaders. Compared to the similar named desktop card it offers reduced clock rates.
High-end professional laptop graphics card based on the consumer GeForce RTX 2070 (TU106 chip) of with 2,304 shaders and 8 GB GDDR6 VRAM. Compared to the normal mobile variant, the Max-Q variants are clocked lower and work in a more efficient state. Currently, we know of three variants with different clock speeds and power consumptions (60, 65 and 70 W).
Efficient high end graphics card with 8 GB GDDR5X-VRAM based on the Pascal GP104 chip. Equipped with 2560 shader units and offers about 10-15% lower performance than a regular GTX 1080 due to lower clocks, but the power consumption is much lower in return.
Pascal based high-end mobile graphics card based on a cut down GP104 (like the GTX 1080) and GDDR5 graphics memory. Successor to the GTX 980M with a 10 Watts higher TDP. Similar to the desktop GTX 1070 performance wise, but with more shaders and lower clock speeds.
Mid range to high end laptop graphics card based on the Turing architecture with raytracing features and 1,920 shaders. Offers reduced clock speeds to achieve a much lower power consumption.
Mid-range graphics card for laptops based on the Ada Lovelace architecture. Offers 6 GB of GDDR6 VRAM (96 Bit) and is configurable with a TGP of 35 - 115W (+ Dynamic Boost).
A lower-end professional graphics card for use in laptops that sports 2,560 CUDA cores and 6 GB of GDDR6 VRAM. It would be fair to say that this is a GeForce RTX 4050 (Laptop) in disguise; consequently, both are powered by the AD107 chip and are fast enough to handle most games at 1080p with quality set to High. The product was brought into existence in February 2024. It leverages TSMC's 5 nm process and the Ada Lovelace architecture; the Nvidia-recommended TGP range is very wide at 35 W to 140 W leading to bizarre performance differences between different systems powered by what is supposed to be the same graphics card.
Mid-range professional laptop graphics card based on the Ampere GA107 chip with 2,560 CUDA cores. Similar to the consumer GeForce GTX 3050 Ti with certified drivers. Available in different performance variants ranging from 35 - 95 W TGP (max. power consumption).
The Intel Arc A730M is a dedicated mobile lower mid-range graphics card based on Xe HPG microarchitecture. It offers 24 cores (384 ALUs), 24 ray tracing units and a base core of 1100 MHz. The TGP is specified between 80 and 120 Watt.
Mid-range mobile graphics card based on the Navi 10 chip (new RDNA architecture) and manufactured in the modern 7nm process. Offers 2,304 cores (36 CUs) and uses fast HBM2 graphics memory.
Mid-range mobile graphics card based on the Navi 10 chip (new RDNA architecture) and manufactured in the modern 7nm process. Offers 2,304 cores (36 CUs) and using GDDR6 graphics memory. The performance should be similar to a RTX 2060 according to AMD.
Mid range dedicated graphics card for laptops based on the desktop GTX 1660 Ti and therefore the Turing architecture without raytracing and Tensor cores. The performance of the Max-P variant should be similar to an older GTX 1070.
Mid-range gaming laptop graphics card based on the Ampere GA107 chip. Offers 2560 shaders and is offered in different TGP (power consumption) variants from 35 to 80 Watt with different clock speeds and therefore performance. Refresh of the RTX 3050 with 6GB, higher shader count but a reduced memory bandwidth of 96Bit.
The Intel Arc A570M is a dedicated mobile lower mid-range graphics card based on Xe HPG microarchitecture. It offers 16 cores (256 ALUs), 16 ray tracing units and a typical gaming core speed of 1300 MHz. The TGP is specified between 75 and 95 Watt.
Mid-range professional laptop graphics card based on the Ampere GA107 chip with 2,048 CUDA cores. Similar to the consumer GeForce GTX 3050 with certified drivers. Available in different performance variants ranging from 35 - 95 W TGP (max. power consumption).
Efficient high end graphics card for thin and light laptops based on the same GP104 chip as the normal GTX 1070 for laptops but at slower clock speeds. Offers a approx. 15% lower performance but much lower power consumption.
The Intel Arc A550M is a dedicated mobile lower mid-range graphics card based on Xe HPG microarchitecture. It offers 16 cores (256 ALUs), 16 ray tracing units and a core speed of 900 MHz. The TGP is specified between 60 and 80 Watt.
The Nvidia Quadro P5000 with Max-Q Design is a mobile high-end workstation graphics card for notebooks. It is the power efficient variant of the normal Quadro P5000 for laptops and offers slightly reduced clock speeds (1101 - 1366 MHz versus 1164 - 1506 MHz) and a greatly reduced power consumption (80 versus 100 Watt TGP).
The most powerful AMD iGPU as of mid 2024. It features 16 RDNA 3.5 architecture CUs (1024 unified shaders) running at up to 2,900 MHz. On average, the 890M finds itself very close to the GTX 1650 Laptop making it possible to play 2024 AAA games at 1080p with most settings set to Low.
Mid-range gaming laptop graphics card based on the new Ada Lovelace AD107 or AD106 chip (like the RTX 4060) but with highly reduced core count (only 1,792 cores) and a small memory bus (64 Bit).
Mid-range gaming laptop graphics card based on the Ampere GA107 chip. Offers 2048 shaders and is offered in different TGP (power consumption) variants from 35 to 80 Watt with different clock speeds and therefore performance. Older variant with 4GB VRAM.
A lower-end professional graphics card for use in laptops that sports 2,048 CUDA cores and 4 GB of GDDR6 VRAM. It appears the RTX 500 is a heavily cut-down GeForce RTX 4050 (Laptop); therefore, we believe the former is powered by the same AD107 chip as the latter. The product was brought into existence in February 2024. It leverages TSMC's 5 nm process and the Ada Lovelace architecture and its gaming performance is sufficient for 2023/2024 games at 1080p on high graphics settings.
The Intel Arc A530M is a dedicated mobile lower mid-range graphics card based on Xe HPG microarchitecture. It offers 12 cores (192 ALUs), 12 ray tracing units and a typical gaming core speed of 1300 MHz. The TGP is specified between 65 and 95 Watt.
Mid-range professional laptop graphics card based on the Ampere GA106 chip with 2,048 CUDA cores. Slower than the consumer GeForce GTX 3050 with certified drivers. Available in different performance variants ranging from 20 - 60 W TGP (max. power consumption).
Entry-level mobile graphics card based on the RDNA 2 architecture. Offers 1024 cores (16 CUs) at a TDP of 50 -80 Watt but only a 64 Bit memory interface.
Entry-level mobile graphics card based on the RDNA 2 architecture. Offers 1024 cores (16 CUs) at a TDP of 35 - 50 Watt but only a 64 Bit memory interface. Low power version of the Radeon RX 6550M with lower clock speeds.
Mid range dedicated graphics card for professional laptops based on the desktop GTX 1650 Ti and therefore the Turing architecture without raytracing and Tensor cores.
The mobile GTX 1060 is based on the GP106 chip and offer 1280 shaders. Compared to the identically named desktop version it features a slightly lower clock rate. Games in Full HD and maximum details should run fluently on the card.
A graphics adapter built into the 9-core and 10-core Apple M4 processors that has direct access to fast on-package LPDDR5x-7500 RAM (120 GB/s bandwidth) as well as hardware support for ray tracing and mesh shading and other modern technologies. It appears its real-world performance is almost as good as the GeForce GTX 1650 Laptop. M4-powered iPad Pros will be much slower in long-term GPU-intensive workloads than M4 iMacs, Mac minis and MacBook Pros due to the lack of proper cooling.
Mid range dedicated graphics card for laptops that uses a Turing TU117 chip without raytracing and Tensor cores. Should be slightly faster than a GTX 1650, but is also available in different variants with a TGP ranging from 55 - 80 Watt (and two packages).
The Intel Arc A370M is a dedicated mobile entry level graphics card based on Xe HPG microarchitecture. It offers 8 cores (128 ALUs), 8 ray tracing units and a base core of 1550 MHz. The TGP is specified between 35 and 50 Watt.
Mid-range mobile graphics card based on the Navi 14 chip (new RDNA architecture) and manufactured in the modern 7nm process. Offers 1,408 cores clocked at 1,645 MHz and using GDDR6 graphics memory.
Mid-range mobile graphics card based on the Navi 14 chip (new RDNA architecture) and manufactured in the modern 7nm process. Offers all 24 CUs (=1,536 shaders) of the chip and therefore 2 more CUs than the similar named mobile Radeon RX 5500M and desktop RX 5500. Available with 4 and 8 GB GDDR6 video RAM.
High-end workstation graphics card based on the 16nm GP104 chip (similar to the Quadro P5000) from the Pascal architecture. The Max-Q version is a power efficient variant of the normal Quadro P4000 with lower clock rates (1113 - 1240 versus 1202 - ?) and a much reduced power consumption (80 versus 100 Watt TGP).
High-end workstation graphics card based on the 16nm GP104 chip using the Pascal architecture. The peak performance is simlar to the older P4000, but the memory bus is cut down to 192 Bit (from 256 Bit). he Quadro GPUs offer certified drivers, which are optimized for stability and performance in professional applications (CAD, DCC, medical, prospection, and visualizing applications).
Polaris 20 based high-end graphics card for laptops, manufactured in an improved 14nm process (LPP+). Should be similar to the Radeon RX 580 and perform similar to a RX 480.
Polaris 20 based high-end graphics card for laptops, manufactured in an improved 14nm process (LPP+). Should be similar to the desktop Radeon RX 580 and perform similar to a RX 480.
Entry-level mobile graphics card based on the RDNA 2 architecture. Offers 768 cores (12 CUs) at a TDP of 25 - 50 Watt and a 64 Bit memory bus for 4 GB GDDR6.
Integrated graphics adapter based on the Xe LPG architecture (similar to the dedicated Arc GPUs, but with a focus on efficiency). Provides all 8 Xe cores (128 Xe vector engines) and 8 ray tracing units.
Mid range laptop graphics card based on the Turing architecture with raytracing features and 2,048 shaders (more than a RTX 2060) but only a small 64 Bit memory bus.
A pretty fast integrated graphics adapter that higher-end Intel Lunar Lake family processors employ. This is a direct successor to the Arc 8; it can drive three SUHD 4320p monitors simultaneously via HDMI 2.1, eDP 1.5 and DP 2.1. With the 140V, all 2023 and 2024 games are playable at 1080p on low graphics settings.
Mid range dedicated graphics card for laptops based on the desktop GTX 1650 and therefore the Turing architecture (TU117 chip in 12nm FFN) without raytracing and Tensor cores. Available with 896 and 1024 shaders.
Mobile mid range graphics card based on the Navi 14 chip (RDNA architecture) that should be slightly faster than a GTX 1650 (if fast GDDR6 VRAM is used).
Efficient high end graphics card for thin and light laptops based on the same GP106 chip as the normal GTX 1060 for laptops but at slower clock speeds. Offers a approx. 15% lower performance but much lower power consumption.
High-end workstation graphics card based on the 16nm GP106 chip (similar to the consumer GTX 1060 Max-Q) from the Pascal architecture. Compared to the normal P3000 offers a reduced power consumption of 60 Watt TGP versus 75 Watt at a slightly reduced performance.
Mid range dedicated graphics card for laptops that uses a Turing TU117 chip without raytracing and Tensor cores. Runs at reduced clock speeds compared to a laptop GTX 1650 Ti to achieve a much lower power consumption (35 versus 50 - 80 Watt).
Mid range dedicated graphics card for laptops based on the desktop GTX 1660 Ti and therefore the Turing architecture without raytracing and Tensor cores. The performance of the Max-P variant should be similar to an older GTX 1070.
Mid range dedicated graphics card for professional laptops based on the consumer GTX 1650 but with less shaders. Therefore it uses the Turing architecture without raytracing and Tensor cores.
An integrated graphics adapter that lower-end Intel Lunar Lake family processors employ. This is a direct successor to the Arc 7 iGPU; DirectX 12 Ultimate is fully supported here and the iGPU also supports hardware-decoding a long list of popular video codecs such as h.266 VVC, h.265 HEVC, h.264 AVC, AV1 and VP9. It is clear the 130V isn't fast enough to run most triple-A 2024 games at 1080p. However, it's still more than good enough for an occasional gaming session Helldivers 2 and Ready or Not run perfectly well at 1080p / Low.
Workstation GPU based on the Polaris architecture with 2304 stream processors (36 CUs) and 8 GB 256 Bit GDDR5 memory. Similar to the consumer and desktop RX 480 / 580 based on Polaris 10/20.
Integrated graphics adapter based on the Xe LPG architecture (similar to the dedicated Arc GPUs, but with a focus on efficiency). Provides 7 of the 8 Xe cores (112 Xe Vector Engines) and 7 ray tracing units.
Most likely based on Polaris 10 based mid-ranged graphics card for laptops. Should be similar to the older Radeon RX 470 / 570 with slightly different clock speeds.
Low-Midrange Graphics Cards - Modern games should be playable with these graphics cards at low settings and resolutions. Casual gamers may be happy with these cards.
The graphics adapter that's built into the smallest Apple M4 series SoC, the one that debuted in October 2024 alongside the Fall 2024 iMac. Its gaming performance is slated to be about as good as that of the AMD Radeon 760M.
Upcoming mobile graphics card based on the Polaris 10 architecture. The GPU is most likely a rebrand of the Radeon RX 470 for laptops with slightly faster clock rates or a higher shader count.
Mid range dedicated graphics card for thin and light laptops based on the GTX 1650 but with reduced clock speeds and power consumption. Uses the Turing architecture (TU117 chip in 12nm FFN) but without Raytracing or Tensor cores.
Mid range dedicated graphics card for professional laptops based on the consumer GTX 1650 Ti. Therefore it uses the Turing architecture without raytracing and Tensor cores.
Mid range dedicated graphics card for professional laptops that offers 768 shaders and therefore less than the GTX 1650. Uses the TU117 chip without raytracing or Tensor cores. The more power-efficient Max-Q variant is using lower clock speeds. Currently two variants with 35 and 40 Watt power consumptions are known.
The Intel Arc A350M is a dedicated mobile entry level graphics card based on Xe HPG microarchitecture. It offers 6 cores (96 ALUs), 6 ray tracing units and a base core of 1150 MHz. The TGP is specified between 25 and 35 Watt.
Mid range dedicated graphics card for professional laptops based on the consumer GTX 1650. Therefore it uses the Turing architecture without raytracing and Tensor cores.
Entry level dedicated graphics card for professional laptops based on the consumer GeForce MX550 with 1024 shaders and a 64 bit memory bus. Therefore it uses the Turing architecture without raytracing and Tensor cores. Sometimes also called Quadro T550.
Integrated graphics card in the slower Kaby Lake-G SoCs. Offers 24 CUs = 1536 shaders at 1063 - 1190 MHz, 4 GB HBM2 graphics memory (on the same package as the GPU and CPU). The TDP is specified at 100 Watt.
High-end workstation graphics card based on the GM204 Maxwell architecture with 1536 shaders and a 256-bit wide memory bus. The GPU is similar to the consumer GTX 980M.
Dedicated graphics card based on the Vega architecture (or a mixture of Polaris and Vega like in the RX Vega M GL). It integrates 20 CUs (1.280 shaders) and 4 GB HBM2 graphics memory.
A fairly powerful RDNA 3.5 architecture iGPU that debuted in June 2024 and a direct successor to the Radeon 780M. Its 12 CUs/WGPs (768 unified shaders) run at up to 2,900 MHz. DX12 Ultimate, ray tracing, AI image generation and other modern features are all supported here, as are many popular video codecs including first and foremost AVC, HEVC, VP9 and AV1. Its gaming performance is good enough for playing 2024 games such as Once Human and The First Descendant at 1080p on low graphics settings.
Integrated graphics unit in the Qualcomm Snapdragon 8 Elite SoC for smartphones with three slices with 1.1 GHz each. Integrated 12 MB shared cache for the three slices. According to Qualcomm, 40% faster than the previous generation (Adreno 750).
A relatively powerful integrated graphics adapter that two Snapdragon X Elite chips feature, as of Sep 2024. It has 1,536 unified shaders running at up to 1,500 MHz. The underlying architecture is reportedly in many respects the same as what was used in the Adreno 730. Performance-wise, the Qualcomm iGPU is almost as good as Intel's Arc 8 in many use cases, allowing the former to deliver playable experience in many 2023 and 2024 games such as Baldur's Gate 3 at 1080p on low graphics settings.
Integrated graphics card in the Ryzen 7040 mobile series APUs based on the RDNA3 architecture with 12 CUs (= 768 shaders) and a clock speed of up to 3 GHz.
An integrated graphics adapter designed by Apple that features 10 cores. This iGPU is built into the Apple M3 SoC; it uses the unified memory architecture and has support for new technologies such as mesh shading and ray tracing. In our testing, the M3 GPU (10 cores) mostly was 5% to 20% faster than the 10-core M2 GPU depending on the task. Just like the rest of the Apple M3 chip, the graphics adapter is manufactured on a 3 nm TSMC process (possibly N3B).
Mainstream graphics card based on Nvidia's Pascal architecture and successor to the GeForce GTX 965M. Is manufactured in a 14nm process at Samsung and the technical specifications are very similar to the desktop version.
Integrated graphics card in the Ryzen 6000 mobile series APUs based on the RDNA2 architecture with 12 CUs (= 768 shaders) and a clock speed of up to 2.4 GHz.
Mid-range workstation graphics card based on the 14nm GP107 chip (similar to the consumer GTX 1050 Ti) from the Pascal architecture. The Quadro GPUs offer certified drivers, which are optimized for stability and performance in professional applications (CAD, DCC, medical, prospection, and visualizing applications).
Mid-range workstation graphics card based on the 14nm GP107 chip (similar to the consumer GTX 1050 Ti) from the Pascal architecture. The Quadro GPUs offer certified drivers, which are optimized for stability and performance in professional applications (CAD, DCC, medical, prospection, and visualizing applications).
High-end workstation graphics card based on the GM204 Maxwell architecture with 1280 shaders and a 256-bit wide memory bus. The GPU is similar to the consumer GTX 970M.
Entry level dedicated graphics card for professional laptops based on the consumer GeForce MX450 with 896 shaders and a 64 bit memory bus. Therefore it uses the Turing architecture without raytracing and Tensor cores. Sometimes also called Quadro T500.
Gimped GeForce GTX 1650 (Turing TU117) based entry level GPU with GDDR5 or GDDR6 graphics memory. Available in 4 different variants, where the LP = low power version with 12 Watt TGP is the slowest.
Dedicated graphics card based on the Vega architecture (or a mixture of Polaris and Vega like in the RX Vega M GL). It integrates 16 CUs (1024 shaders) and 4 GB HBM2 graphics memory.
Integrated graphics card in the slower Kaby Lake-G SoCs. Offers 20 CUs = 1280 shaders at 931 - 1011 MHz, 4 GB HBM2 graphics memory (on the same package as the GPU and CPU). The TDP is specified at 65 Watt.
Integrated professional graphics card in the slower Kaby Lake-G SoCs. Offers 20 CUs = 1280 shaders at 931 - 1011 MHz, 4 GB HBM2 graphics memory (on the same package as the GPU and CPU). The TDP is specified at 65 Watt. Technically identical to the Radeon RX Vega M GL but with support for professional drivers.
High-end workstation graphics card based on the GM204 Maxwell architecture with 1024 shaders and a 256-bit wide memory bus. The GPU is similar to the consumer GTX 965M.
Mainstream graphics card based on Nvidia's Pascal architecture and successor to the GeForce GTX 960M. Is manufactured in a 14nm process at Samsung and the technical specifications are very similar to the desktop version.
Efficient mid-range graphics card for thin and light laptops based on the same GP107 chip as the normal GTX 1050 for laptops but at slower clock speeds. Offers a approx. 10 - 15% lower performance than the laptop GTX 1050 but at a much lower power consumption.
A fairly powerful RDNA 3.5 architecture iGPU that debuted in early 2025 and a direct successor to the Radeon 760M. Its 8 CUs/WGPs (512 unified shaders) run at up to 3 GHz. DX12 Ultimate, ray tracing, AI image generation and other modern features are all supported here, as are many popular video codecs including first and foremost AVC, HEVC, VP9 and AV1. Its gaming performance is good enough for playing 2024 games in low graphics settings.
Integrated graphics card offering 8 of the 10 cores designed by Apple and integrated in the Apple M3 SoC. Offers a new architecture with dynamic caching, mesh shading and ray tracing support.
Integrated graphics card offering 10 cores (depending on the model) designed by Apple and integrated in the Apple M2 SoC. According to Apple 25% faster with a slightly higher power consumption compared to the M1 GPU with 8 cores. The TDP is rated at 15W, in our tests the GPU needed 13.5W under load.
Integrated graphics card in the Ryzen 7040 mobile series APUs based on the RDNA3 architecture with 8 of the 12 CUs (= 512 shaders) and a clock speed of up to 2.8 GHz.
An integrated graphics adapter that four Qualcomm Snapdragon X processors feature, as of Sep 2024. It has 1,536 unified shaders at its disposal that are said to run at up to 1,250 MHz. The underlying architecture is reportedly in many respects the same as what was used in the Adreno 730. While thankfully faster than the aging Intel Iris Xe (96 EUs), the Adreno fails to match the Radeon 780M meaning Baldur's Gate 3, a triple-A title released in 2023, is barely playable at 1080p on low with 23 fps on average. Triple-A games of 2024 fare even worse, in no small part due to Qualcomm's subpar graphics drivers.
Integrated graphics adapter in the custom CPU of the Steam Deck OLED with 8 CUs and a clock rate of 1000 - 1600 MHz. Manufactured at TSMC in the N6 process (6nm).
Integrated graphics card in the Ryzen 6000 mobile series APUs based on the RDNA2 architecture with 6 of the 12 CUs (= 384 shaders) and a clock speed of up to 1.9 GHz.
Workstation GPU based on the Polaris architecture with 896 stream processors (14 CUs) and 4GB 128 Bit GDDR5 memory. Similar to the consumer and desktop RX 460 / 560 based on Polaris 11.
Lower high-end graphics card based on the Maxwell chip GM204 and GM206 (newer 2016 version with higher clock rates). Part of the GeForce GTX 900M series.
Mid-range workstation graphics card based on the 14nm GP107 chip (similar to the consumer GTX 1050 Ti) from the Pascal architecture. The Quadro GPUs offer certified drivers, which are optimized for stability and performance in professional applications (CAD, DCC, medical, prospection, and visualizing applications).
Mobile mid-range to high-end graphics card from AMD that is most likely based on the old Pitcairn chip with 1024 shaders and used in the mid-range 2015 iMac 5k.
The Intel Iris Xe MAX (DG1) is a dedicated mobile entry level graphics card based on the Gen 12 architecture (similar to the integrated graphics of the Tiger Lake processors). Compared to the Iris Xe iGPU (e.g. in the i7-1185G7), the Xe MAX offers 4 GB LPDDR4x graphics memory and higher boost clock speeds (up to 1.65 GHz versus 1.35 GHz).
Mobile GPU based on Polaris 11 that can be ordered as an option for the more powerful Apple MacBook Pro 15-inch Late 2016. Should be a similar to a Radeon RX 460, but with more shaders at a lower clock speed.
Integrated graphics card in Intel Tiger Lake G4 SoCs based on the new Gen. 12 architecture with 96 EUs (Execution Units / Shader Cluster). The clock rate depends on the processor model. The Tiger Lake chips are produced in the modern 10nm+ process at Intel.
Mainstream workstation graphics card based on the GM107 Maxwell architecture with 640 shaders and a 128-bit wide memory bus. The GPU is similar to the consumer GTX 960M.
Pascal based entry level workstation graphics card with certified drivers for stability and performance in professional applications. Most likely based on the GeForce MX150.
Polaris based entry level dedicated graphics card with 512 shaders (8 CUs) or 640 shaders (10 CUs) and 4 GB GDDR5 graphics memory (64 bit memory bus). Compared to the older RX 550X, the memory bus is halfed.
Entry level workstation graphics card based on the Polaris 12 chip like the consumer counterpart Radeon RX 550. Uses 640 shaders (10 compute units) and a 128 Bit GDDR5 memory system (1000 MHz).
Graphics card for the Apple MacBook Pro 15 2018 based on Polaris 21. The specifications are identical to the previous Radeon Pro 455 / 555 with slightly higher clock speeds and 4 instead of 2 GB GDDR5 graphics memory.
Mobile GPU based on Polaris 11 that is found in the more powerful Apple MacBook Pro 15-inch Late 2016. Probably a slimmed-down version of the Radeon RX 460 (clocks and possibly shader count).
Pascal based entry level workstation graphics card with certified drivers for stability and performance in professional applications. Most likely based on the GeForce MX150.
Successor of the MX150 and still based on the same Pascal GP108 chip (similar to the desktop GT 1030) but with higher clock speeds. Available in two versions, a normal 25 Watt version and a low power version with 10 Watt TDP and reduced performance. The 25 Watt version e.g. offers a 21% higher boost clock than the old MX150 leading to a 5% performance gain.
Successor of the MX230 and most likely a renamed GeForce MX250/MX150 based on the Pascal GP108 chip with 384 shaders and a 64 Bit memory bus for GDDR5.
Pascal GP108 based laptop graphics card and mobile version of the desktop GeForce GT 1030. Offers 384 shader cores and usually 2 GB GDDR5 with a 64 Bit memory bus. Manufactured in 14nm.
Integrated graphics card in Intel Tiger Lake G7 SoCs based on the new Gen. 12 architecture with 80 EUs (Execution Units / Shader Cluster). The clock rate depends on the processor model. The Tiger Lake chips are produced in the modern 10nm+ process at Intel.
A fairly powerful RDNA 3.5 architecture iGPU that debuted in early 2025 and a direct successor to the Radeon 760M. Its 8 CUs/WGPs (512 unified shaders) run at up to 3 GHz. DX12 Ultimate, ray tracing, AI image generation and other modern features are all supported here, as are many popular video codecs including first and foremost AVC, HEVC, VP9 and AV1. Its gaming performance is good enough for playing 2024 games in low graphics settings.
Integrated graphics card in the Ryzen 7040 mobile series APUs based on the RDNA3 architecture with 4 of the 12 CUs (= 256 shaders) and a clock speed of up to 2.5 GHz.
Integrated graphics card based on the Xe LPG architecture (similar to the dedicated Arc GPUs, but with a focus on efficiency). Provides 4 Xe cores (64 Xe vector engines) and 4 ray tracing units.
Mainstream workstation graphics card based on the GM107 Maxwell architecture with 640 shaders and a 128-bit wide memory bus. The GPU is similar to the consumer GTX 960M/950M.
Polaris based mid-range graphics card with 512 shaders (8 compute units) and a maximum clock speed of 1,219 MHz (according to AMD). Very similar to the older RX 540 with maybe slightly higher clock speeds.
Entry level workstation graphics card based on the Polaris 12 chip like the consumer counterpart Radeon RX 540X. Uses 512 shaders (8 compute units) and a 128 Bit GDDR5 memory system.
Polaris based mid-range graphics card with 512 shaders (8 compute units) and a maximum clock speed of 1,219 MHz (according to AMD). Uses the new Polaris 12 chip like the faster Radeon RX 550 (not verified but highly likely) with the slightly improved 14nm FinFET process.
Entry level workstation graphics card based on the Polaris 12 chip like the consumer counterpart Radeon RX 540. Uses 512 shaders (8 compute units) and a 64 Bit GDDR5 memory system (96 GB/s).
A relatively fast graphics adapter that the Apple A18 Pro SoC comes equipped with. This iGPU has 6 cores and is RT-enabled; it appears it is based on the same architecture as M4-series GPUs. It managed to beat the 7-core Apple M1 GPU in many tests in our October 2024 testing.
Mobile GPU based on Polaris 11 that is found in the entry-level Apple MacBook Pro 15-inch Late 2016. Probably a slimmed-down version of the Radeon RX 460 (clocks and possibly shader count).
An integrated graphics adapter that the Snapdragon X Plus 8-core X1P-46-100 SoC features. Compared to the faster 3.8 TFLOPS and 4.6 TFLOPS X1-85 iGPUs, this one does not just run at lower clock speeds but also has fewer unified shaders at its disposal, with 768 being the most likely number. The underlying architecture is reportedly not much different from what was used in the Adreno 730. As for its gaming performance, it's only just sufficient for pre-2021 games at resolutions such as HD 720p on low graphics settings.
Entry level graphics card with 512 shaders, 4 GB GDDR5 and a 64 Bit memory bus. Could be based on the Polaris architecture (although the RX is missing in the name). Therefore, it could be a rebrand of the RX 540.
Entry level graphics card with 512 shaders, 2 GB GDDR5 and a 64 Bit memory bus. According to AMD, the 540X uses the 4th gen of GCN. Compared to the RX 540X, the memory bus is cut down and the pixel fillrate is greatly reduced.
Pascal based entry level workstation graphics card with certified drivers for stability and performance in professional applications. Most likely based on the GeForce MX150.
Mid-range workstation graphics card based on the 28nm GM107 Maxwell architecture with 512 shaders and a 128-Bit wide memory bus. Similar to the older Quadro M1000M and consumer GeForce GTX 950M (which features more shaders).
Pascal based entry level workstation graphics card with certified drivers for stability and performance in professional applications. Most likely based on the GeForce MX150 with a reduced shader count.
Integrated graphics card in Intel Alder Lake SoCs based on the new Gen. 12 architecture (Xe) with 64 EUs (Execution Units / Shader Cluster). The clock rate depends on the processor model.
Integrated graphics card in Intel Ice-Lake G7 SoCs based on the new Gen. 11 architecture with 64 EUs (Execution Units / Shader Cluster). The clock rate depends on the processor model ranging from 300 MHz base to 1050 - 1100 MHz boost. The Ice-Lake chips are produced in the modern 10nm process at Intel produced.
An integrated graphics adapter that the Snapdragon X Plus 8-core X1P-42-100 SoC features. Compared to the faster 3.8 TFLOPS and 4.6 TFLOPS X1-85 iGPUs, this one does not just run at lower clock speeds but also has fewer unified shaders at its disposal, with 768 being the most likely number. The underlying architecture is reportedly not much different from what was used in the Adreno 730. As for its gaming performance, it's only just sufficient for pre-2020 games at resolutions such as HD 720p on low graphics settings.
Mid-range mobile graphics card that features 640 shader cores (10 compute cores) and a 128 Bit GDDR 5 memory controller (2 GB GDDR5 in the Apple MacBook Pro 15).
Integrated graphics card in Intel Tiger Lake G4 SoCs based on the new Gen. 12 architecture with 48 EUs (Execution Units / Shader Cluster). The clock rate depends on the processor model. The Tiger Lake chips are produced in the modern 10nm+ process at Intel.
Integrated graphics card in Intel Tiger Lake-h SoCs based on the new Gen. 12 architecture with 32 EUs (Execution Units / Shader Cluster). The clock speed depends on the processor model.
Integrated graphics card offering 8 of the 10 cores designed by Apple and integrated in the Apple M2 SoC. Offers a performance situated between the old 8 core GPU in the Apple M1 and the 10-core version in the higher end models. The power consumption is around 10 Watt under load.
Integrated graphics card offering 8 cores (depending on the model) designed by Apple and integrated in the Apple M1 SoC. According to Apple faster and more energy efficient as competing products (like the Tiger Lake Xe GPU).
Integrated graphics card offering 7 cores (of the 8 in the chip) designed by Apple and integrated in the Apple M1 SoC. According to Apple faster and more energy efficient as competing products (like the Tiger Lake Xe GPU).
Mid-range laptop graphics card based on the Maxwell architecture. Compared to the older 940M, the MX now also supports GDDR5 graphics memory and maybe slightly faster clock speeds.
Mid-range workstation graphics card based on the 28nm GM108 chip with 384 shaders and a 64 Bit memory bus. Most likely based on the consumer Geforce 940MX with GDDR5 graphics memory.
The Apple A18 GPU integrates five cores, hardware ray tracing and mesh support. Apple states that the A18 GPU is up to 40% faster than the A16 GPU in the iPhone 15. In our benchmarks, the graphics unit is positioned between the old A16 GPU and the A17 Pro GPU with 6 cores.
Integrated graphics card in Intel Ice-Lake G4 SoCs based on the new Gen. 11 architecture with 48 of the 64 EUs (Execution Units / Shader Cluster). The clock rate depends on the processor model ranging from 300 MHz base to 1050 - 1100 MHz boost. The Ice-Lake chips are produced in the modern 10nm+ process at Intel produced.
Integrated graphics card offering 8 cores designed by Apple and integrated in the A12Z Bionic SoC. Compared to the A12X Bionic in the iPad Pro 2018 models (7 cores) and A12 Bionic (iPhone XS series) it should offer the same architecture with more cores.
Mainstream workstation graphics card based on the GM107 Maxwell architecture with 384 shaders and a 128-Bit wide memory bus (as opposed to the GM108-based K620M with a 64-Bit memory bus).
Graphics chip for smartphones and tablets that is integrated within the Qualcomm Snapdragon 8 Gen 3 SoC. Qualcomm claims that it is 25% faster than the Adreno 7340 in the Snapdragon 8 Gen 2 and first benchmarks show that the GPU can beat the Apple A17 Pro iGPU.
Graphics chip for smartphones and tablets that is integrated within the Qualcomm Snapdragon 8 Gen 2 SoC. Qualcomm claims that it is 25% faster than the Adreno 730 in the Snapdragon 8 Gen 1 and first benchmarks show that the GPU can beat the Apple A16 iGPU (previously the fastest smartphone GPU).
A fast integrated GPU for ARM processors that has hardware ray tracing, Accuracy Super Resolution and other modern technologies. It's slated to be almost as fast as the Qualcomm Adreno 830.
High-end graphics chip for smartphones with 12 of the 16 possible cores. Based on the new "5th Gen" architecture (successor to Valhall) including Variable Rate Shading (VRS) and Ray Racing. Compared to the G715, 15% higher performance should be possible at the same clock rates.
The ARM Immortalis-G715MC11 is an integrated high-end graphics card for smartphones with 11 of the 16 possible cores. It is based on the third generation of the Valhall architecture by ARM and offers a 15% higher performance than the previous generation. The G715 is the first mobile GPU with hardware ray tracing support and also supports variable rate shading (VRS).
High-end graphics chip for smartphones with 7 of the 16 possible cores. Based on the new "5th Gen" architecture (successor to Valhall) including Variable Rate Shading (VRS) and Ray Racing. Compared to the G715, 15% higher performance should be possible at the same clock rates.
The Qualcomm Adreno 732 is an integrated graphics unit that is used in the Snapdragon 7+ Gen 3. The performance is slightly above the older Adreno 730.
Graphics chip for smartphones and tablets that is integrated within the Qualcomm Snapdragon 8 Gen 1 SoC. Qualcomm claims that it is 30% faster than the Adreno 660 in the Snapdragon 888 SoC. For Android it should be one of the fastest graphics cards in late 2021.
The ARM Immortalis-G715MC11 is an integrated high-end graphics card for smartphones with 11 of the 16 possible cores. It is based on the third generation of the Valhall architecture by ARM and offers a 15% higher performance than the previous generation. The G715 is the first mobile GPU with hardware ray tracing support and also supports variable rate shading (VRS).
Mid-range graphics chip for smartphones with 6 of the 6 possible cores. Based on the fourth generation of Valhall architecture. Offers variable rate shading.
Graphics chip for smartphones and tablets, integrated in the Qualcomm Snapdragon 7+ Gen 2 SoC. According to Qualcomm, 2x faster than the old Adreno 644 in the Snapdragon 7 Gen 1.
The ARM Mali-G710MC10 is an integrated high-end graphics card for smartphones with 10 of the 16 possible cores. It is based on the third generation of the Valhall architecture by ARM which offers a 20% higher performance and energy efficiency than the previous generation. In the Dimensity 9000, the G710 is produced in the modern 4nm process.
Workstation graphics card based on the 28 nm GM108 Maxwell chip with 384 shaders and a 64-bit memory interface. Compared to the faster M600M, the M500M features only DDR3 graphics memory with a 64-bit bus and is therefore similar to the old Quadro K620M (also Maxwell GM108).
Graphics chip for smartphones and tablets, integrated in the HiSilicon Kirin 9020. Unfortunately, Huawei does not provide any detailed information. What is certain is that it does not support hardware-based ray tracing and only the older Vulkan 1.0 standard.
Graphics chip for smartphones and tablets that is integrated within the Samsung Exynos 2200 SoC. Based on the AMD RDNA 2 architecture with 3 CUs (=192 shaders). Manufactured in the modern 4nm process (EUV).
Graphics chip for smartphones and tablets that is integrated within the Qualcomm Snapdragon 888 SoC. Qualcomm claims that it is 35% faster than the Adreno 650 in the Snapdragon 865 SoC and offers a 20% better power efficiency.For Android it should be one of the fastest graphics cards in 2021.
Integrated graphics card based on the 2nd generation of the Valhall architecture. Uses all of the 24 possible clusters. The Mali-G78MP24 is one of the fastest mobile GPUs for Android based phones in 2020.
The ARM Mali-G710MC10 is an integrated high-end graphics card for smartphones with 10 of the 16 possible cores. It is based on the third generation of the Valhall architecture by ARM which offers a 20% higher performance and energy efficiency than the previous generation. In the Dimensity 9000, the G710 is produced in the modern 4nm process.
Integrated graphics card based on the 2nd generation of the Valhall architecture. Uses 22 of the 24 possible clusters. The Mali-G78MP22 is one of the fastest mobile GPUs for Android based phones in 2020.
Entry class dedicated graphics card with 320 or 384 shaders (depending on the used chip) and DDR3 or GDDR5 graphics memory (both with a 64 Bit bus). Both chips are used for a few years, e.g. even starting with the Radeon 8590M (28nm GCN).
Integrated graphics card designed by Apple with 20% more performance than the previous generation (A12). Includes 4 cores and uses 40% less power according to Apple.
Graphics chip for smartphones and tablets that is integrated within the Qualcomm Snapdragon 865 SoC. Qualcomm claims that it is 25% faster than the Adreno 640 in the Snapdragon 855 SoC and offers 50% more compute units (ALUs). For Android it should be one of the fastest graphics cards in 2020.
Integrated graphics card based on the 2nd generation of the Valhall architecture. Uses 20 of the 24 possible clusters. The Mali-G78MP20 is one of the fastest mobile GPUs for Android based phones in 2020.
Integrated graphics card based on the 2nd generation of the Valhall architecture. Uses 14 of the 24 possible clusters. The Mali-G78MP14 is a fast mobile GPUs for Android based phones in 2020.
The Apple A10X Fusion / PowerVR graphics card is integrated in the Apple A10X Fusion SoC (benchmarks and specs) that can be found in the Apple iPad Pro from 2017. Most likely based on PowerVR technology. According to Apple it integrates 12 cores (compared to the 6 cores of the A10 GPU) and is therefore significantly faster.
Integrated graphics card in Intel Tiger Lake-h SoCs based on the new Gen. 12 architecture with 24 of the 32 EUs (Execution Units / Shader Cluster). The clock speed depends on the processor model. Only suited for very low demanding games.
Dual graphics combination of a integrated Radeon R5 (Bristol Ridge) and a dedicated Radeon R7 M440. Suffers from micro stuttering and in some games even slower than the R7 M440 alone.
Entry level graphics card with 320 shaders (5 compute units) or 384 shaders (6 compute units), 4 GB DDR3 or GDDR5 and a 64 Bit memory bus. Therefore, comparable to an old Radeon R7 M445 / M440 / M340 / M255 (28nm GCN).
Entry level dedicated graphics card based on a 28nm GCN (2.0 / 1.2?) chip. According to our sources may be slower than an R5 M330 and M335 (slower clock speed and same amount of compute cores).
Entry level graphics card with 320 shaders (5 compute units), 2 GB GDDR5 and a 64 Bit memory bus. Uses the same chip as the old Radeon R7 M230, M340, M440 and e.g. Radeon R7 M445 (28nm GCN).
A pretty basic integrated graphics adapter sporting 2 RDNA 2 CUs (128 unified shaders) running at a clock speed of up to 2.2 GHz. Its performance is close to what the Vega 6 and the 32 EU Iris Xe deliver.
Integrated graphics card in Intel Ice-Lake G1 SoCs based on the new Gen. 11 architecture with 32 of the 64 EUs (Execution Units / Shader Cluster). The clock rate depends on the processor model ranging from 300 MHz base to 1050 - 1100 MHz boost. The Ice-Lake chips are produced in the modern 10nm process at Intel produced.
Integrated graphics card in Intel Tiger Lake-H SoCs based on the new Gen. 12 architecture with 16 of the 32 EUs (Execution Units / Shader Cluster) or 128 Shaders. The clock rate depends on the processor model.
Integrated graphics card in the Tegra X1 SoC without dedicated memory. The X1 GPU is based on the Maxwell architecture with two SMMs (256 shader cores). It supports DirectX 11.2, OpenGL ES 3.1 and OpenGL 4.5 (same features as the desktop and laptop Maxwell based graphics cards).
Integrated graphics card in the Bristol Ridge APUs and based on the third generation GCN architecture with 384 shader cores. Clock speeds and performance depend on the clock speed (720 to 800 MHz) configurable TDP limit (15 to 35 Watt versions) and main memory.
Integrated GPU (GT2, 24 EUs) found on some Kaby-Lake-Refresh CPU models (15 W ULV series). Technically identical to the previous Kaby-Lake GPU called HD Graphics 620.
Integrated graphics card in the Qualcomm Snapdragon 8cx Gen 3 SoC for Windows laptops. According to Qualcomm 60% faster than the previous generation. Not very well suited for 3D games due to missing driver support and low performance.
Integrated graphics card in the Microsoft SQ1 ARM SoC. Similar to the Adreno 680 in the Qualcomm Snapdragon 8cx SoC for Windows laptops but higher clocked and therefore with a higher theoretical performance (2 versus 1.8 TFlops).
Integrated graphics card in the Qualcomm Snapdragon 8cx SoC for Windows laptops. According to Qualcomm 2x faster than the previous Adreno 630 in the Snapdragon 850 for Windows PCs with a 60% improved efficiency (thanks to the 7nm process).
Integrated graphics adapter of some AMD-A dual-core APUs ("Stoney Ridge"). Based on the GCN architecture, implements 192 shader units and runs at up to 800 MHz. The performance can vary heavily depending on the TDP (15-25 Watts for the whole chip)
Multicore graphics card included in the Apple A10 Fusion SoC that can be found in the Apple iPhone 7 and 7 Plus smartphones. Most likely based on PowerVR technology and according to apple 40% faster than the A9 while using only 2/3 of the power.
Integrated graphics card based on the Valhall architecture. Uses 11 of the 16 possible clusters and according to ARM offers 30% better performance at 30% better efficiency compared to a Mali-G76.
Integrated graphics card based on the Valhall architecture. Uses 9 of the 16 possible clusters and according to ARM offers 30% better performance at 30% better efficiency compared to a Mali-G76.
Integrated graphics card based on the second generation of the Bifrost architecture. Uses 16 of the 20 possible clusters and according to ARM offers twice the performance per cluster than the old Mali-G72. It is used in the Kirin 990 5G and according to Huawei should be about 6% faster and 20% more efficient than a Adreno 640 in a Snapdragon 855.
Graphics chip for smartphones and tablets that is integrated within the Qualcomm Snapdragon 7 Gen 1 SoC. Qualcomm claims that it is 20% faster than the previous generation.
Graphics chip for smartphones and tablets that is integrated within the Qualcomm Snapdragon 780G SoC. Qualcomm claims that it is 50% faster than the Adreno 620 in the Snapdragon 768G SoC.
Mid-range graphics chip for smartphones with 2 of the 6 possible cores. Based on the fourth generation of Valhall architecture. Offers variable rate shading.
Graphics chip for smartphones and tablets that is integrated within the Qualcomm Snapdragon 778G SoC. Qualcomm claims that it is 40% faster than the Adreno 620.
Graphics chip for smartphones and tablets, integrated in the Qualcomm Snapdragon 7s Gen 2 SoC. According to Qualcomm, 2x faster than the old Adreno 644 in the Snapdragon 7 Gen 1.
Graphics chip for smartphones and tablets that is intergrated within the Qualcomm Snapdragon 855 SoC. Qualcomm claims that it is 20% faster than the Adreno 630 in the Snapdragon 845 SoC and offers 50% more compute units (ALUs).
Graphics chip for smartphones and tablets, integrated in the HiSilicon Kirin 9000W. Unfortunately, Huawei has not provided any detailed information. What is certain is that it does not support hardware-based ray tracing and only the older Vulkan 1.0 standard.
Integrated graphics card in the Qualcomm Snapdragon 845 SoC. According to Qualcomm 30% faster than the old Adreno 540 in the Snapdragon 835 with 30% less power consumption.
Integrated graphics card based on the second generation of the Bifrost architecture. Uses 12 of the 20 possible clusters and according to ARM offers twice the performance per cluster than the old Mali-G72.
Integrated graphics card based on the second generation of the Bifrost architecture. Uses 10 of the 20 possible clusters and according to ARM offers twice the performance per cluster than the old Mali-G72.
Integrated graphics card in the Qualcomm Snapdragon 835 SoC. Slighly optimized architecture compared to the Adreno 530 but with higher clock speed due to the new 10 nm process. In the beginning of 2017 one of the fastest graphics cards for Android based smartphones.
Integrated graphics card with 18 cores (from 32) clocked at up to 850 MHz (according to ARM). The G72MP18 uses the second generation of Bifrost architecture and supports modern APIs like OpenGL ES 3.2, Vulkan 1.0, OpenCL 2.0, DirectX 12 FL11_1 and Renderscript.
Integrated graphics card in Intel Jasper Lake Pentium Silver models based on the Gen. 11 architecture. Offers all 32 EUs (Execution Units / Shader Cluster). The clock rate depends on the processor model ranging from 350 MHz base to 900 MHz boost. The Jasper Lake chips are produced in the 10nm process at Intel produced (similar to Ice Lake).
Integrated graphics card in some Amber Lake Y-series processors (7 Watt TDP) with 24 EUs (GT2) and no dedicated graphics memory. Clocks with up to 1050 MHz depending on the CPU Model.
Integrated graphics card in some Amber Lake Y-series processors (5 Watt TDP) with 24 EUs (GT2) and no dedicated graphics memory. Clocks with up to 1000 MHz depending on the CPU Model.
Integrated graphics card in Intel Jasper Lake Celeron mid range models based on the Gen. 11 architecture. Offers 24 of the 32 EUs (Execution Units / Shader Cluster). The clock rate depends on the processor model ranging from 350 MHz base to 900 MHz boost. The Jasper Lake chips are produced in the 10nm process at Intel produced (similar to Ice Lake).
Integrated graphics card in some Kaby Lake Y-series processors (4.5 Watt TDP) with 24 EUs (GT2) and no dedicated graphics memory. Clocks with Update to 1050 MHz depending on the CPU Model.
Middle class graphics card for smartphones which is integrated in ARM SoCs. Supports OpenGL ES 3.2, Vulkan 1.1, OpenCL 1.2 and Android NN HAL. The performance should be slightly better than a Adreno 430.
Integrated graphics card in the Qualcomm Snapdragon 820 that is clocked at up to 624 MHz. Early 2016 it is a high end graphics card for Android based smartphones and tablets that should be able to coop with very demanding games.
In Tegra K1 SoC integrated graphics card based on the Kepler architecture. Features a single SMX with 192 cores and full OpenGL 4.4 support (e.g. with Tessellation).
Integrated GPU consisting of 6 clusters with 192 ALU (FP32) and support for Metal/OpenGL ES 3.x. The GT7600 is the fastest smartphone GPU at the time of announcement on September 2015.
Integrated graphics card based on the second generation of the Bifrost architecture. Uses 4 of the 20 possible clusters and according to ARM offers twice the performance per cluster than the old Mali-G72.
Integrated GPU (GT1) with 12 EUs found on some low-end Pentium Gold CPUs (e.g. Whisky-Lake based 15 W ULV series). Same performance as the old HD Graphics 610 and 510.
Integrated graphics card in Intel Jasper Lake Celeron entry level models based on the Gen. 11 architecture. Offers 16 of the 32 EUs (Execution Units / Shader Cluster). The clock rate depends on the processor model ranging from 350 MHz base to 800 MHz boost. The Jasper Lake chips are produced in the 10nm process at Intel produced (similar to Ice Lake).
Integrated low-end graphics adapter with DirectX 12 support, which can be found in some ULV SoCs from the Gemini Lake series. Compared to the HD Graphics 505, the 605 offers improved display outputs.
Integrated GPU that can be found e.g. in the Mediatek Helio P23 SoC with 700 or 770 MHz clock speed. Supports OpenGL ES 3.2, Vulkan 1.0, OpenCL 2.0, DirectX 11 FL11_1 and RenderScript.
Integrated graphics adapter of some AMD-A dual-core APUs ("Stoney Ridge"). Based on the GCN architecture, implements 192 shader units and runs at up to 600 MHz depending on the model. The performance can vary heavily depending on the TDP (10-15 Watts for the whole chip).
Integrated smartphone and tablet GPU that supports OpenGL ES 3.1 and DirectX 11.2 (FL 11_1). The graphics card has 192 unified shaders clocked at up to 650 MHz.
Integrated graphics adapter of some AMD-E2 APUs ("Stoney Ridge"). Based on the GCN architecture, implements 128 shader units and runs at up to 600 MHz depending on the model. The performance can vary heavily depending on the TDP (10-15 Watts for the whole chip).
The Broadcom VideoCore 7 is an integrated GPU with 12 cores and up to 800 MHz clock frequency. It is installed, for example, in the Broadcom BCM2712 SoC (Raspberry Pi 5) and is manufactured here in 16nm.
Integrated mid range graphics card with 6 clusters. ARM states that the G52 series is 30% faster and 15% more efficient compared to the G51 series on the same process node.
Integrated mid range graphics card with 2 clusters. ARM states that the G52 series is 30% faster and 15% more efficient compared to the G51 series on the same process node.
Integrated mid range graphics card with 1 clusters. ARM states that the G52 series is 30% faster and 15% more efficient compared to the G51 series on the same process node. The MC1 / MP1 variant is the smallest version.