Benchmarking the FirePro V7900, V5900 and V4900 / The V4900 shined on SPECapc Lightwave
Posted by Alex Herrera on December 11th 2011 | Discuss
Categories:
Hardware Review
Tags:
nvidia
amd
market
graphics
intel
firepro
The FirePro V7900 and V5900 boards, based on AMD’s Caymen GPUs, now occupy the mid-range and high-end positions in AMD’s workstation graphics portfolio. The V4900 workstation AIB is the latest addition to the line. It’s built with the Turks Northern Island’s GPU, and targeted at the entry level. AMD provided us with all three new models to assess their competitive positioning.
To get a sense of the performance the new trio can deliver, we employed some of the same basic tools we have in the past, including one relatively new (and very welcome) addition. We benchmarked the three boards shortly after they were released, but not simultaneously. We tested the V5900 and V7900 in August 2011 and the V4900 in November 2011.
We tapped SPEC’s Viewperf to isolate the stress on the graphics boards, and we use SPECapc, Cadalyst and Cinebench tests to get a handle on how well the boards support whole-system performance. Choosing among the SPECapc suite of benchmarks, we chose Lightwave and 3ds Max, representing two popular applications used primarily in digital content creation (DCC).
The latter was especially useful, as the latest version, SPECapc for 3ds Max 2011 has been updated. Improvements include new models and test cases, a cleaner and enhanced user interface, and most important, compatibility with more recent versions of both 3ds Max (2011) and Windows (both 32 and 64-bit Windows 7).
| CPU | Intel Xeon X5670, 6-core @ 2.93 GHz |
| Memory | 12 GB of 1333 MHz DDR3 |
| Disk | 2 x 500 GB (7200 rpm SATA) |
| OS | Windows 7, 64-bit |
Our Z800 test bench
Thanks to HP, we’ve got an impressive test bench for professional graphics boards. The company was kind enough to equip us with their top-end Z800, capable of driving up to four PCI Express boards at once. With a 6-core Intel Xeon X5670 CPU (Westmere generation) running at 2.93 GHz, paired with 12 GB of 1333 MHz DDR3 memory, our Z800 provides a powerful and consistent foundation upon which to assess the performance of these and other workstation-caliber boards.
Viewperf 11
The first test was with Viewperf 11, the most recent revision of the long-time graphics benchmark. A nice step up from Viewperf 10, this version thankfully eliminated the multitude of confusing options on running the benchmark and in their place is a simple dialog box that allows you to select resolution and number of iterations. Simple. Thank you, SPEC.
Where 1280 x 1024 used to be a reasonable common denominator for setting resolution, we’re now pushing to 1920 x 1080, as the resolution is now generally pervasive and supported by most moderately priced monitors.
Not surprisingly, the three boards delivered results commensurate with their GPU specifications and price tags: the V7900 bested the V5900 by a significant margin, and ditto for the latter over the V4900. The V7900 managed a 31% margin over the V5900 on one viewset (ensight-04) but performed not materially better on two other viewsets (lightwave-01 and proe-05). On average, the V7900 was just around 10% faster than its sister mid-range board. Likewise, V5900 outperformed the entry V4900 in similar fashion, with a big gap on ensight-04 (64.6%) and more modest deltas elsewhere.
Similarly, the results for the V5900 and V7900 compared to other similarly priced boards with results submitted to SPEC (e.g. Nvidia’s Quadro 2000 and Quadro 4000, respectively) also varied widely. With some viewsets (e.g. lightwave), the boards all performed similarly. However, in some viewsets (e.g. esight, tvcis) the AMD boards bested their rivals substantially, while in others (e.g. catia, proe), they lagged far behind. While difficult to compare overall, it appears the AMD boards are on relatively equal footing with their rival’s price-comparable SKUs.
| 3 iterations | V7900 perf. edge over V5900 | V5900 perf. edge over V4900 |
|---|---|---|
| catia-03 | 9.2% | 13.6% |
| ensight-04 | 31.0% | 64.6% |
| lightwave-01 | -0.2% | 6.6% |
| maya-03 | 13.6% | 9.2% |
| proe-05 | 0.2% | 2.2% |
| sw-02 | 3.4% | 4.8% |
| tcvis-02 | 15.1% | 1.0% |
| snx-01 | 7.1% | 28.6% |
SPECapc: Lightwave and 3ds Max 2011
To evaluate overall system performance, we employed two SPECapc tests, for 3ds Max and Lightwave, along with Cadalyst 2011 and Cinebench. While 3ds Max is increasingly used as a visualization tool in CAD spaces like architecture, engineering and construction (AEC), both Lightwave and 3ds Max are used in digital content creation.
However, with SPECapc, neither the graphics nor the rest of the system are solely responsible for scores. Rather, it’s the sum of the parts that are being tested, to try to give the user an idea of how the system might perform in a real-world environment. Where GPU-accelerated rendering is the bottleneck, throughput will vary by graphics AIB. But where it’s not (CPU, memory or I/O limitations), the test won’t show a difference. Average it all out, and the differences in results for whole-system type tests like these come in lower than for Viewperf.
The V4900 shined on SPECapc Lightwave, as it performed on par with its higher-priced siblings and actually nudged out the V5900 on all three sub-tests. That parity needs to be taken in context. First off, SPECapc Lightwave is notoriously CPU-bound, especially for the Render and Multi-task sub-tests of Lightwave (ironic for something named “Render”, but in this case rendering is performed in software on the CPU, not via hardware-accelerated APIs like OpenGL or DirectX) … so no big surprise that the results for all three boards were nearly identical.
SPECapc Lightwave’s Interactive does more GPU-based rendering and the V7900 did show an edge over the V5900 (4.1% faster).
We like the way SPEC organized the sub-test composite scores for 3ds Max 2011, separating out GPU and CPU.
The overall Large Model Composite is a weighted blend of test results for two sequences: the CPU creating a large scene, and the GPU viewing the scene (with varying viewpoints/ports).
As we’d expect, with SPECapc 3ds Max’s GPU Composite (+10.4%) and Large Model Composite (+13.3%), the V7900 has the clear edge, not as dramatic an edge as it has in some Viewperf tests, but an edge nonetheless.
On the other hand, the V4900’s Large Model Composite score was a head-scratcher. Since the V4900 is a lower-end board with fewer resources than either of its two bigger siblings, we would expect to see the Large Model Composite score somewhat reduced, but not to the degree we observed. Looking at the individual scores that comprise the composite, we noticed that it wasn’t just the GPU portion that was dramatically reduced, but the CPU component as well. The latter result made no sense, as all three boards were running on the same test-bench.
Cadalyst
The CAD space represents the largest chunk of workstation and professional graphics users. But while 3ds Max gets a share of work in CAD as well as DCC, there’s one application that dominates the former space: AutoCAD. And when it comes to getting a feel on how well a board or system can handle AutoCAD, we fortunately still have the mainstay Cadalyst benchmark, now updated to 2011.
Results from the Cadalyst 2011 tests are in perfect agreement with the boards’ pricing and positioning. The more the test relies on the GPU, the bigger the performance delta (+6.5% for 3D Graphics and +2.4% for Total) as we step up from new Northern Islands’ FirePro SKUs. And the less the test or sequence relies on GPU horsepower — that is, execution ends up being CPU or memory bound — the less of a difference measured (+0.4% for 2D).
Cinebench
Last but not least, we have Cinebench. Cinebench renders a 3D sequence (currently, a car chase through town) utilizing both CPU and GPU resources. It reports a composite performance number for both CPU (CPU) and graphics (OpenGL). Since we’re benchmarking graphics boards, and comparing two boards on the same Z800 test bench, we focus on the OpenGL component.
Once again, we see that the inexpensive V4900 shines. Running the V5900 on the new driver would have been very interesting, and it’s something we’ll probably get a chance to do in the context of AMD’s next board release.
Of interest to both a graphics board supplier and its customers is how products in the vendor’s portfolio are positioned with respect to one another.
Commensurate with its higher price, the V7900 shows consistent performance superiority over the V5900 on SPEC tests, but less so on Cadalyst (a fact AMD’s likely fine with, considering that market segment is not the target for the V7900). Viewing these system-level results in combination with Viewperf 11, we could justifiably (albeit roughly) peg the V7900 as a 10% better overall performer, with the potential to deliver perhaps 20-30% better performance in localized instances.
By contrast, the V5900 had a tougher time separating itself from its lesser sibling, the V4900, which given its price and lesser GPU, was a pleasant surprise. But given that different drivers were involved, we need to take that observation with a large grain of salt.
| Benchmark | Test / Metric | V7900 perf. edge over V5900 | V5900 perf. edge over V4900* |
|---|---|---|---|
| SPECapc for 3ds Max 2011 | Large Model | 13.3% | 282.5%** |
| GPU | 10.4% | 0.2% | |
| SPECapc for Lightwave 9.6 | Interactive | 4.1% | -0.7% |
| 3D Graphics Index | 6.5% | 4.7% | |
| Total index | 2.4% | 3.5% | |
| Cinebench | OpenGL | 12.7% | -4.8% |
What do we think?
The graphics arm of AMD, what we once called ATI, is clearly on the rebound. Looking back, we see the timely release of the company’s Evergreen GPU generation back in September of 2009 as the watershed moment marking AMD’s resurgence. Evergreen drew a very competitive line in the sand, perhaps not delivering the all-out max performance that Nvidia’s subsequent Fermi generation did, but establishing some compelling price/performance.
The company followed up on its hallmark performance-per-dollar with Northern Islands, a generation that made sensible and relatively low-risk improvements over Evergreen. And the Northern Islands GPU most capable and appropriate for workstation use is Cayman. In the FirePro V7900 and V5900 boards, we see the first workstation-caliber usage for Cayman, and the results are what we expected. Both offer compelling price/performance, and the V7900 manages a reasonable-but-not-overwhelming advantage over the V5900.
While Cayman is the premiere Northern Islands’ GPU, don’t count out other members of the family. As the first and only FirePro board to tap the Turks GPU, the V4900 demonstrated some very compelling price performance. The exact extent of the V4900’s prowess relative to the V5900 is not completely known, as we had that third variable of a newer driver to consider when weighing the results. But regardless, the V4900’s numbers still say an awful lot about the price/performance of the Turks GPU. Its impressive performance particularly on Cadalyst dovetails perfectly with its market positioning. Both AMD and Nvidia tend to push the lower half of their product lines more toward that AutoCAD user base.
After several years in which Nvidia executed almost flawlessly—while ATI had the stumbling block of its acquisition by AMD to deal with—AMD’s share had fallen dramatically in the professional graphics market, down to a low of only 8.6% unit share by mid 2008. Riding on the back of Evergreen, the company’s managed to boost that share to 17.1% in the first quarter of 2011, and move it a nudge higher in Q2 to 18.5%, its highest level since the close of 2006.
Doubling share is of course a commendable feat, but it would have put a far bigger dent in rival Nvidia’s share had it been doubling from its former levels (e.g. 27.7% in Q1’05), not its all-time low. Now Northern Islands is here bearing the FirePro brand, and based on the price/performance merits of the initial trio of boards, we don’t imagine AMD will be conceding any significant unit share in the near term. And given the company has already signed on workstation leaders Dell, HP and Fujitsu to drive OEM sales for the V4900, we expect market leader Nvidia will be feeling noticeably more pressure than it’s been accustomed to. – A.H.
> > Reviewing the HP Z210 Small Form Factor (SFF) workstation
< < Review of ASUS MARS II 3GB: the fastest AIB in the history of the world
