In the present day, we wished to point out off some fascinating QCT QuantaGrid options for pc engineering. Earlier this week, STH confirmed the efficiency of Intel Xeon Max for plenty of totally different workloads. Final week, Patrick, our Chief Analyst was in Taipei and visited Quanta and noticed a few of these options firsthand. We had plenty of questions on “why Intel Xeon Max” and since we did each of those items just lately, we figured we might get Quanta to get us an instance, and that instance is the QCT for automotive.
QCT QuantaGrid for Intel Xeon MAX in Automotive CFD and CAE
Simply setting the stage, we wished to point out how the Xeon MAX (or different processors/ accelerators) can be utilized for computer-aided engineering and design. QCT despatched us this slide for servers it makes use of when it builds automotive trade options which have heavy CFD and CAE parts of the workflow.
We truly noticed the QCT QuantaGrid D54Q-2U in Taipei this final week. Since we’re utilizing belongings from that journey (QCT sponsored us flying to Taipei) we’re going to say that is sponsored by QCT. Though not our regular studio high quality photographs, right here is the server that makes use of EDSFF NVMe SSDs on the entrance:
Right here is an summary of the 2U server:
Here’s what the CPU sockets appear to be. These are essential as a result of they’ll settle for as much as 350W 4th Gen Intel Xeon Scalable CPUs and Xeon MAX CPUs. There are some techniques on the market that don’t help the upper TDP energy supply.
One of many huge features of the server is the airflow that is ready to cool these high-end CPUs nonetheless utilizing air cooling. We will even present liquid-cooled outcomes quickly for Xeon MAX.
Since we had simply gone over among the Xeon MAX versus Xeon Platinum outcomes earlier this week, and QCT is ready to use each within the server, the corporate had examples of what occurs with Xeon MAX. One of many huge instances was doing knowledge visualization for CFD utilizing OpenFOAM, so we requested about what the variations have been utilizing OpenFOAM and Xeon MAX versus 4th Gen Intel Xeon Scalable.
Here’s what QCT got here up with utilizing the Intel Xeon Platinum 8458P and the Intel Xeon MAX 9468 in HBM-only mode with each SNC=1 and SNC=4 modes. Except for that, we have to level out on this comparability that the Intel Xeon Platinum 8458P is a 44-core half whereas the Xeon MAX 9468 is a 48-core half. Whereas there are about 9% extra cores, clock speeds on the Xeon Max are 8-22% decrease than the Xeon Platinum being proven right here. The larger takeaway is admittedly the SNC tuning.
We simply seemed on the Intel Xeon Max 9480 CPU, the 56-core model in our Intel Xeon MAX 9480 Deep-Dive if you wish to study extra about that and what “HBM-only” and SNC=4 means.
It was nice seeing different outcomes that highlighted the purpose we made in our Intel Xeon MAX piece that it may be considerably sooner than commonplace Xeons in some workloads, but additionally could be tuned to extend efficiency much more. In case of us surprise, a part of the thought is that for some high-value components of the design pipeline, one can merely swap Xeon MAX CPUs into the identical servers and deploy widespread platforms throughout a corporation. Once more, QCT despatched us the Automotive trade for example displaying the HPC components and the opposite components of the pipeline that is likely to be serviced by different components.
That was one thing that we realized final week at QCT that they’re truly driving these options and selecting CPUs and even GPU-based techniques based mostly on the functions getting used at totally different components of the workflow.
QCT additionally has a DevCloud that features plenty of totally different instruments and workflows starting from functions like OpenFOAM to AI and storage options.
We now have a bit coming that we recorded in Taiwan about liquid cooling Intel Xeon MAX servers and the event platforms, so keep tuned for that.
Closing Phrases
We wished to take this chance to point out off an awesome use case for the Intel Xeon MAX CPUs that we confirmed earlier this week and the way these work in total workflows. Since we have been at QCT final week, we have been capable of get them to ship an instance from an automotive workflow the place QCT normally profiles totally different components of the pipeline after which picks the correct servers and CPUs for these servers. Some have seen our protection and thought that going Xeon MAX was all or nothing. It looks as if how these are literally getting deployed is extra like what one would count on, in focused components of the pipeline.
At STH many people are automotive lovers as effectively and normally automotive design is a good instance of the place HPC clusters are used for among the engineering workflows, however then different techniques are used at different steps. It appeared like a straightforward approach to mix the 2 themes from the primary half of September right into a extra sensible instance. One query we now have gotten is “the place is that this used?” and we wished to shot this for example.
