Dell Precision 7960 And 5860 Review: Powerful Workstations Up To 56 Cores
Dell Precision 7960 & 5860 Workstations: System And Pro-Vis Graphics Benchmarks
ATTO Disk Benchmark
The ATTO disk benchmark is a quick and simple test which measures read/write bandwidth and IOPS across a range of different data sizes. While we don't typically compare these results across multiple machines, it's useful to gauge whether a particular system's storage subsystem can handle being put under pressure.
Both of these Dell workstations are configured with RAID 0 volumes which stripe data across two drives to improve performance. This allows both PCIe 4 drive arrays to achieve 10GB/s read speeds rivaling PCIe Gen 5 drives, but that is not to say the arrays perform identically. The Precision 7960 delivers better read throughput at mid-to-low transfer sizes and boasts significantly higher write throughput across the board, even exceeding the read performance. This disparity is most apparent when looking at IOPS where the difference is as much as a factor of 10. Actual storage performance will, of course, vary by workload and storage configuration, so it is an important consideration as you spec out one of these or similar systems.
Cinebench R23
Cinebench rendering performance test based on Cinema 4D from Maxon. Cinema 4D is a 3D rendering and animation tool suite used by animation houses and producers like Sony Animation and many others. It's very demanding on system processor resources and can utilize up to 256 threads, which makes it an excellent gauge of computational throughput. This is a multi-threaded, multi-processor aware benchmark that renders and animates 3D scenes and tracks the length of the entire process. The rate at which each test system was able to render the entire scene is represented in the graphs below.
The Dell Precision 7960 outpaces the HP Z8 Fury G5 in multi-threaded performance, but it still isn’t enough to beat out the last-gen Threadripper Pro 5995WX in the Precision 7865. Unfortunately for the Xeon w9-3495X, the TR Pro 5995WX has recently been supplanted by the Threadripper Pro 7995WX with a whopping 96 cores of compute. We don’t have figures to compare directly here, but check out our early performance review of the new Zen 4 chip to see its awesome potential.
The Dell Precision 5860’s Xeon w7-2495X is dwarfed by any of the titans above, but it is not trying to compete for the podium, either. This 24-core chip provides about 35% better per-core performance scaling than the 56-core Xeon x9-3495X while operating at a much lower TDP.
LAME XP Audio Encoding
For this test, we created 16-copies of our own 223MB WAV file (an 11 minute Grateful Dead jam) and converted it to the MP3 format using the multi-thread capable LAME XP tool. Processing times are recorded below, listed in minutes:seconds. Shorter times equate to better performance.
Lame XP’s multi-core scaling is limited to a maximum of 32 threads, which makes it difficult to differentiate a winner here, but the Intel chips do at least perform favorably in comparison to AMD’s Zen 3 flagship.
7-Zip Data Decompression
The 7-Zip benchmark measures decompression performance using the LZMA method, which leverages the Lempel–Ziv–Markov chain algorithm to perform lossless data compression. The benchmark produces a final rating in GIPS (giga instructions per second).
The Dell Precision 7960 once again edges out the HP Z8 Fury G5 in 7-Zip, although the Xeon W flagship has fallen further behind AMD’s last-gen flagship. If pure compute performance is the primary consideration then the Threadripper Pro camp is probably the one to be in, but it does come with a pricing premium that may keep Intel in the game.
POV-Ray CPU Ray Tracing Benchmark
POV-Ray, or the Persistence of Vision Ray-Tracer, is an open source tool for creating realistically lit images. We tested with POV-Ray's standard 'one-CPU' and 'all-CPU' benchmarking tools on all of our test machines and recorded the scores reported for each. Results are measured in pixels-per-second throughput; higher scores equate to better performance.
POV-Ray cements the performance hierarchy we’ve seen in previous tests, which goes to show that Dell is extracting about all that can be done from the Xeon w9-3495X. Likewise, the Precision 5860 is performing well for its weight class, though the HP has retained better single-threaded performance than either Dell by a skosh across these workloads.
Y-Cruncher Multi-Threaded Pi Benchmark
Y-Cruncher is a multi-core-capable tool that calculates the value of pi to a specified number of digits. In this case, we ran the tool on all threads and had the application calculate the value of pi to 1 billion digits. The values below are the time required to perform the calculation expressed in seconds. As a result, lower values indicate better performance.
Y-Cruncher throws a wrench in that performance hierarchy, though. The HP Z8 Fury G5 emerges victorious this time, with the Threadripper Pro 5995WX equipped Precision 7865 falling to third place. This benchmark is quite memory intensive, so the Precision 7960’s quad-channel configuration is unable to keep pace with the eight-channel Z8 Fury G5. We highly recommend maxing out memory channels when configuring your system so as to not leave performance on the table.
UL 3DMark CPU Physics Benchmark
For the 3DMark Physics test, we simply create a custom 3DMark run consisting solely of the physics test, which is multi-threaded and CPU dependent, and report the results.
None of our contenders come out looking great here. These kinds of gaming-adjacent workloads are best left to consumer desktop processors that can push high clocks with low inter-core latencies.
BlackMagic RAW Video Encoding Speed
The Blackmagic RAW Speed Test is a CPU and GPU benchmarking tool that tests full-resolution Blackmagic RAW video decode performance. The tool can be used to evaluate the performance at various resolutions and bitrates on the CPU or using OpenCL on a GPU. We're reporting two results here, both using 8K resolution, but at differing bitrates and compression levels.
At 12:1 compression, the Dell Precision 7960 and HP Z8 Fury G5 are virtually deadlocked. However, moving to the less compressed 3:1 workload tips the scales in favor of the HP machine’s eight-channel memory configuration.
The Precision 5860 is notably not too far behind which is respectable for its much lower system cost, and we can see diminishing returns in effect for this use case.
After truly impressive results from our Precision 7865 and Z8 Fury G5 reviews, we find today’s scores rather underwhelming at first glance. For starters, the Z8 Fury G5 used a quad-GPU configuration with eight-channel memory that allowed Blackmagic RAW decode performance to soar.
Sure, the 12:1 results posted by the Precision 7960 are the best of any single GPU we have seen, but the RTX A6000 equipped Precision 5860 is only 20 frames per second behind.
The Precision 7960 is most negatively impacted by its quad-channel memory configuration, but we know this can be alleviated by a full 8-channel configuration. However, the Precision 5860’s Xeon W-2400 series chips only support up to quad-channel memory, so that limitation could be a consideration for some if they are using memory bandwidth restricted workloads like this.
Blender Rendering Benchmarks
Blender is a free and open-source 3D creation suite that can handle everything from modeling, rigging, and animation to simulation, rendering, compositing, and motion tracking. It has a purpose-built benchmarking tool that will track the time it takes to complete rendering a particular model (or models). We used both the CPU and GPU focused benchmarks with all three models currently available in the benchmark tool.
Performance scales well for both of these Dell Precision workstations with CPU rendering. The Precision 7960 beats out the Z8 Fury G5 in all three renders while bearing the same Xeon W chip. Meanwhile, the Precision 5860’s Xeon w7-2495X delivers a bit over half the performance of the Xeon w9-3495X with less than half the core count (24 cores vs 56 cores).
The Precision 7960 is particularly impressive in the GPU test. We’ve left in some reference consumer GPU results for comparison here and find the RTX 6000 Ada goes toe to toe with the NVIDIA GeForce RTX 4090 in all cases. That is all while drawing much less power (300W vs 450W TGP) and occupying a much smaller formfactor.
Conversely, the Precision 5860’s result weighs in lower than we expected. The Precision 7865 churned out a few hundred more samples per minute in each test with the same RTX A6000 GPU. It’s better than the HP Z8 Fury G5’s showing, but that result came with a few caveats to explain it.
Running workloads directly within Blender shows the Dell Precision 7960’s prowess. Its render times are equivalent to three RTX A6000’s running within the HP Z8 Fury G5 and not far off the full quad A6000 configuration. RTX 6000 Ada is an uncontested Blender rendering beast in the professional GPU arena.
The Precision 5860’s result is competitive here within its weight class, but the discrepancy between its RTX A6000 and the RTX 6000 Ada Generation card is stark.
SPECviewperf 2020 v3 OpenGL / DX Benchmarks
SPECviewperf includes a variety of tests, which produce significantly different framerates, so we’ve sorted them into three groups to make the results a bit easier to sort through. Please note the legends at the bottom of each chart, which designate the application or viewset used…
The RTX 6000 Ada-equipped Precision 7960 surges ahead in most of the workloads tested in SPEC, yielding up to nearly double the performance of the RTX A6000 contenders. This advantage isn’t so pronounced for all tests though. For instance, creo-03, catia-06, and maya-06 are hardly distinguished.
The Precision 5860’s ranking varies from result to result, but it is probably fair to say that certain tests like 3Dsmax-07 and snx-04 are not its best showings.
