Overclock.net banner
1 - 4 of 4 Posts

185 Posts
Discussion Starter · #1 ·
A fair amount of confusion exists when considering professional and desktop video cards, at least for those relatively new to the subject. It can be a minefield for some, and I hope this thread helps those who are seeking some relief. During my investigation I have encountered a mountain of misinformation, with much conflicting advice provided. For example, "Don't waste your money on a professional card when a gaming card is the same thing and will perform just fine." The inverse is also quite common: "Stop wasting your money on gaming cards when a professional adapter can handle both work and play equally well."

Instead of posting pure generality and supposition, I would prefer measurements and hard numbers, especially when making big assertions. As Carl Sagan once famously quipped, "extraordinary claims require extraordinary evidence." Yes, indeed. With this in mind, I will post my comparative analysis between two hard-hitting platforms: a server/workstation using an nVidia Quadro K5000 professional graphics adapter for professional 3D modeling, and another powerful desktop featuring Radeon HD 7970 adapters in CrossFireX for occasional gaming.

Along the way, I hope the information in this thread evolves to produce a useful trade-off analysis for those interested in gaining a broader perspective on this subject.

185 Posts
Discussion Starter · #3 ·
Quadro K5000 Versus Radeon HD 7970 in CrossFireX

Two strong platforms are involved in this analysis. One machine features the nVidia Quadro K5000 professional graphics adapter, based on the Kepler architecture. The other is configured in CrossFireX with two Radeon HD 7970 cards. Besides a clear difference in video processors, both machines are otherwise very similar in terms of CPU, memory and I/O capacity. There should be no issue with CPU or memory bottlenecks when evaluating either video platform.

Quadro K5000 Server/Workstation Specficiations

Cooler Master HAF XB Case
ASUS Z9PA-D8 dual LGA 2011 ATX mainboard with PIKE 2008 daughter RAID adapter
KingWin LZP-1000, 80 PLUS Platinum Modular Power Supply
2 x Intel Xeon E5-2670 8-Core 2.6GHz Sandy Bridge-EP Processor, 3.3GHz Turbo Boost, 20MB L3 Cache
64GB 8 x Samsung 8GB DDR3 1600 MHz ECC Registered Server Memory M393B1K70DH0-CK0
nVidia Quadro K5000 Graphics Card 1536 CUDA cores 4GB GDDR5 Memory VCQK5000-PBN ($1,750)

Radeon HD 7970 CrossFireX Desktop/Workstation Specficiations

Cooler Master Cosmos II Case
Enermax MaxRevo 1350W Power Supply
MSI Big Bang-XPower II Intel LGA 2011 X79 Mainboard
Intel Core i7-3960X Extreme Edition Sandy Bridge-E 3.3GHz LGA 2011 CPU
Corsair Dominator GT 32GB (4 x 8GB) DDR3 1866Mhz (PC3 15000) 1.5v CAS 9-10-9-27 SDRAM
Two XFX Double Dissipation Black Edition Radeon HD 7970 in CrossFireX ($860)

Desktop/Gaming Benchmarks

To evaluate high-stress 3D gaming load, benchmarks for both platforms were gathered using FurMark v1.10.4 at 1920x1080 resolution. All video cards were running at stock speed (no overclocking).

The Quadro K5000 system scored 2,011 points, averaging 33 FPS with a 60C maximum GPU temperature. The Radeon HD 7970 CrossFireX system scored 6,313 points, averaging 105 FPS with a 68C maximum GPU temperature.

The chart and graph reveals how the Radeon hardware is far better suited for gaming workloads:

Based on these results, it is clear the gaming card performed much better under heavy gaming load, providng three times the performance for half the cost. How far behind is the Quadro K5000 in terms of gaming experience? The gaming score produced by the Quadro K5000 platform is roughly equivalent to prior-generation AMD Radeon HD 5870 or NVIDIA GeForce GTX 580 gaming cards.

Professional Benchmarks

To evaluate heavy professional modeling load, benchmarks for both platforms were gathered using SPECviewperf 11 at 1600x1200 resolution.

The Quadro K5000 system scored an average of 73.14 and the Radeon HD 7970 CrossFireX system averaged 21.13. The chart and graph reveals how the Quadro hardware is far superior when rendering large 3D models. In fact, the tables turn completely, with the professional card providing well over three times the performance on average. Moreover, several vendor-specific (CAD and Maya especially) tests are six to ten times slower when running the gaming card:


These numbers present a clear distinction between professional and gaming video hardware: Both video platforms do well within their intended domain, and both perform poorly when loaded outside their domain. Therefore, for those who require good results under heavy professional and gaming loads, the answer is clearly two distinct systems. However, this clear distinction can begin to blur somewhat when the heavy loading factor is reduced from very stressful conditions down to much lighter workloads. This is the point where that confusion mentioned at the top of the thread introduces itself. Additional analysis is needed with reduced loading factors to determine when or if either video platform becomes viable working outside their intended domain.
1 - 4 of 4 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.