Review iBuyPower Valkyrie CZ-15 Notebook

Case quality is very similar to that of the MSI GT60 and GX60. The reseller does include some customized finishing to differentiate it from competing DTRs, most notably the web-like glossy and matte texture and design on the brushed metal outer lid and palm rests. This contrasts very well to the smooth gloss and brushed surface of the larger CZ-17, though the admittedly stylish blue streaks on the larger model are now gone. The textured surface unfortunately attracts fingerprints much more easily than its larger brother and can be even more difficult to wipe off compared to any grease or grime on the keyboard. Still, for a chassis that is over a year old, it holds up quite well against the more common Clevo models and feels thicker and more rigid than the similarly sized Asus G55 ROG models.

Keyboard

Even though the case is smaller than the CZ-17, the keyboard is only 1 cm shorter both in length and width. This is essentially negligible, though it would have been nice to have a larger Enter key and dedicated volume/brightness toggles. The top row of touch-based commands makes a return and users will need to press on each one firmly or the input may not register at all.

Touchpad

The touchpad is still absurdly small at about 7.75 x 4.75 cm. Inputs respond reliably, but multi-touch is inherently difficult. The click keys are also too shallow in depth and with somewhat disappointing feedback. Fortunately, most users will be using an external mouse for all intents and purposes and the touchpad will even automatically disable once a USB mouse is detected. However, with other gaming notebooks like the Asus G750JW starting to use much larger touchpads, MSI and iBuyPower may have to eventually follow suit.

Color space reproduction covers much of the sRGB spectrum but only a little more than half of the AdobeRGB standard at 57 percent. This is very similar to the recently released MSI GT70H and even the CZ-17, so this is a non-issue for gamers. Digital artists who demand the most stringent color reproduction for professional work may have to look elsewhere.

Further analyses with the CalMAN 5 software reveal excellent color accuracy with an average DeltaE of 3.83. In general, colors are more accurate as they become more saturated with green being the best on the CZ-15. Grayscale is good as well and is better than our tested Clevo W350ST, Eurocom X3 and Clevo W650SR, though not up to par with the tested Asus G750JW.

CPU performance in synthetic benchmarks is similar to other notebooks in our database equipped with the same processor and in most cases is right on the center. When compared to the Ivy Bridge i7-3740QM that it supplants, single-threaded performance in Super Pi is more or less identical between the two processors while multi-thread wPrime results are actually better on the older IVB CPU (247 seconds vs. 273 seconds). This may simply be due to the higher raw clock speed advantage that the i7-3740QM has over the i7-4700QM. This is confirmed by CineBench R11.5 where the Haswell CPU in the CZ-15 is nearly identical in the single-core benchmark to the i7-3740QM (1.52 points vs. 1.53 points), but is noticeably slower in the multi-core benchmark (6.76 points vs. 6.93 points).

The CZ-15 can be configured with up to a GTX 780M. Our base model is one step down with the GTX 770M, which offers much less CUDA cores (960 vs. 1344) compared to the previous generation GTX 680M and narrower memory bandwidth (96.2 GB/s vs. 115.2 GB/s). Nonetheless, performance is sufficient for gaming in native 1080p with more reserved graphical settings recommended if playing advanced titles. On the CZ-15, the GPU will idle at 135/202.5 MHz core/memory with a rated base speed of 811 MHz. Turbo was observed to be as high as 823 MHz during our stress test.

Of the tested titles, all were able to play well above 30 FPS on high 720p settings. Some titles, such as Sleeping Dogs, Metro Last Light and Company of Heroes 2 are enough to be unplayable at the highest of settings at full HD. This is compared to the GTX 780M, which can at least maintain 30 FPS on a couple of these titles even on max settings at 1080p. The GTX 680M does show similar performance compared to the 770M, but certain titles may actually run smoother on the former due in part to the larger memory interface.

More benchmarks and comparisons can be seen in our full review of the Kepler refresh here.

We perform our standard stress test with Prime95 and FurMark to stress the CPU and GPU, respectively, with the same monitoring tools we utilized during our CZ-17 stress test. With only Prime95 active, the CPU was observed to be stable at its rated 3.2 GHz Turbo Boost for the first minute, but quickly settled on an operating range between 2.9 GHz and 3.1 GHz, which is 100 MHz to 300 MHz less than its Turbo Boost maximum for four active cores. As such, the CPU was unable to maintain maximum CPU Turbo Boost after a few minutes of full CPU stress. According to HWiNFO, CPU temperature tops out at about 85 degrees C under these conditions.

With only FurMark active, the Nvidia GPU was observed to be operating at its 823 MHz turbo speed. However, after the first minute or so, clock speed dropped to a stable 796.9 MHz to 810 MHz range according to GPU-Z, which is only a hair below its rated base clock. Fortunately, memory was stable throughout our test at 1002.4 MHz. Note that fan noise was observed to be increasing much more quickly with a fully active GPU than a fully active CPU. 

With both Prime95 and FurMark running simultaneously for over an hour, both the CPU and GPU were able to maintain 2.9 GHz to 3.1 GHz and 796.9 MHz, respectively, without any hitches. Both processors benefit from short-lived Turbo Boosts that cannot be maintained under full or independent stress. No throttling was observed, though we note that CPU and GPU temperature reached a high 95 degrees C and 86 degrees C, respectively. A 3DMark 11 run immediately following the test resulted in no significant reductions in final scores and thus no throttling.

If running on batteries, the CPU becomes unaffected, but the discrete GPU will take a very large hit in performance as it maxes out at only 405/405 MHz for the core/memory even if on the High Performance Windows profile. The stunted performance is confirmed by running 3DMark 11 on batteries, which returned essentially the same Physics score (6626 points on AC power vs. 6254 points on batteries) and less than half the GPU score (4856 points vs. 2309 points on batteries). 

The 7800 mAh, 87 Wh removable Li-Ion battery is the same as the battery module of the GX60 and older GT60. As a result, runtime comparisons between these similar, yet uniquely equipped notebooks can reveal a lot about the power savings between the various CPU and GPU models used.

Between the CZ-15, GX60 and GT60, not too large of a discrepancy in WLAN runtimes can be observed. All three notebooks are expected to last about 4.5 hours of constant use, though the AMD-powered GX60 may last for around half an hour longer on average. Most interesting is that the idling runtime of the CZ-15 is at least a few hours longer than those MSI notebooks mentioned. For example, whereas the GT60 with its Ivy Bridge and GTX 6xxM hardware can last for about 6 hours when idling, the CZ-15 can idle for around 9.5 hours. This may be partly attributed to the higher 1200 MHz idling clock rate of many Ivy Bridge processors versus just 800 MHz of idling Haswell processors in addition to other minor software and hardware optimizations.

In the end, the longer idling runtime should allow users to squeeze out more life from the CZ-15 by lowering system settings accordingly.