by Rob Williams on July 26, 2017 in Graphics & Displays
AMD last month rounded out its Radeon Pro WX series with the WX 3100 and WX 2100. These nearly identical cards differ themselves by their framebuffer: 4GB vs. 2GB. It’s the beefier of the two cards we’re taking a look at here – a low-profile option that sips little power, and sits quiet while it works away.
On the previous page, I mentioned that SPEC is an organization that crafts some of the best, most comprehensive benchmarks going, and in a similar vein, I can compliment SiSoftware. This is a company that thrives on offering support for certain technologies before those technologies are even available to the consumer. In that regard, its Sandra benchmark might seem a little bleeding-edge, but at the same time, its tests are established, refined, and accurate across multiple runs.
While Sandra offers a huge number of benchmarks, just four of the GPU ones are focused on: Cryptography, Financial Analysis, Scientific Analysis, and also memory bandwidth. Some of the results are a bit too complex for a graph, so a handful of tables are coming your way.
Cryptography
Regardless of your encryption needs, the WX 3100 is going to deliver extremely good performance for its price point. NVIDIA has an enormous gain in the hashing scheme of things, but the WX 7100 kills it on the encrypt/decrypt side. What’s up with the WX 5100 coming below the WX 4100, you ask? It’s a bit of an odd duck: it performs great overall, and just where I’d expect it to, but in certain tests, it lags behind. Either an architectural gotcha, or a sign that the card should have a power connector.
Financial Analysis
|
Sandra 2017 – Financial Analysis (FP32) |
|
Black-Scholes |
Binomial |
Monte Carlo |
NVIDIA Quadro P4000 |
6.5 G/s |
1.1 M/s |
2.9 M/s |
AMD Radeon Pro WX 7100 |
5.2 G/s |
1.3 M/s |
1.9 M/s |
NVIDIA Quadro P2000 |
3.8 G/s |
656 k/s |
1.8 M/s |
AMD Radeon Pro WX 5100 |
3.4 G/s |
478 k/s |
672 k/s |
AMD Radeon Pro WX 4100 |
2.2 G/s |
531 k/s |
773 k/s |
AMD Radeon Pro WX 3100 |
2.5 G/s |
321 k/s |
467 k/s |
|
Sandra 2017 – Financial Analysis (FP64) |
|
Black-Scholes |
Binomial |
Monte Carlo |
NVIDIA Quadro P4000 |
622 M/s |
63 k/s |
129 k/s |
AMD Radeon Pro WX 7100 |
958 M/s |
81 k/s |
239 k/s |
NVIDIA Quadro P2000 |
360 M/s |
36 k/s |
75 k/s |
AMD Radeon Pro WX 5100 |
406 M/s |
49 k/s |
97 k/s |
AMD Radeon Pro WX 4100 |
395 M/s |
35 k/s |
98 k/s |
AMD Radeon Pro WX 3100 |
219 M/s |
18 k/s |
55 k/s |
SPECwpc on the previous page showed absolutely zero deviation between GPUs in its financial test, but Sandra proves that scaling is just fine as long as the workloads can take advantage of them. NVIDIA’s got the Monte Carlo game locked down, but the Radeon Pros are extremely competitive with the other algorithms.
Scientific Analysis
|
Sandra 2017 – Scientific Analysis (FP32) |
|
GEMM |
FFT |
N-Body |
NVIDIA Quadro P4000 |
3.1 TFLOPS |
128 GFLOPS |
2.7 TFLOPS |
AMD Radeon Pro WX 7100 |
2.5 TFLOPS |
210 GFLOPS |
2.2 TFLOPS |
NVIDIA Quadro P2000 |
1.8 TFLOPS |
87 GFLOPS |
1.7 TFLOPS |
AMD Radeon Pro WX 5100 |
945 GFLOPS |
138 GFLOPS |
755 GFLOPS |
AMD Radeon Pro WX 4100 |
1 TFLOPS |
85 GFLOPS |
917 GFLOPS |
AMD Radeon Pro WX 3100 |
670 GFLOPS |
70 GFLOPS |
647 GFLOPS |
|
Sandra 2017 – Scientific Analysis (FP64) |
|
GEMM |
FFT |
N-Body |
NVIDIA Quadro P4000 |
156 GFLOPS |
95 GFLOPS |
127 GFLOPS |
AMD Radeon Pro WX 7100 |
276 GFLOPS |
80 GFLOPS |
194 GFLOPS |
NVIDIA Quadro P2000 |
90 GFLOPS |
53 GFLOPS |
84 GFLOPS |
AMD Radeon Pro WX 5100 |
123 GFLOPS |
56 GFLOPS |
103 GFLOPS |
AMD Radeon Pro WX 4100 |
109 GFLOPS |
33 GFLOPS |
84 GFLOPS |
AMD Radeon Pro WX 3100 |
63 GFLOPS |
33 GFLOPS |
49 GFLOPS |
Once again, AMD exhibits strong performance with the WX 3100 here, performing quite close in parts to its bigger brother, the WX 4100.
Bandwidth
The WX 3100 performs very well against its big bro in this bandwidth test, and while these values don’t align with vendor-provided ones, they scale much the same way.