i9-7980XE vs. Xeon Gold 6146 floating point

transpower

Honorable
Feb 12, 2014
24
0
10,510
I've searched but cannot a floating point calculation comparison between the Intel i9-7980XE and Xeon Gold 6146. Does anyone know of an article about this subject?
 
Solution
The Xeon Gold 6146 has 2 AVX 512 units per core, which makes them faster for floating point operations that utilise Intel's Advanced Vector eXtensions.

However, putting that aside, since you use it for highly critical business tasks, reliability is key. The i9 7980XE may tend to overload current X299 boards, and may have some very slight instability issues, which are unacceptable in your use case.

In this case, Xeon 6146 is for you as it supports ECC RAM (error correction ram), more RAM (up to 1TB vs 128GB) and is also binned for 24/7 usage, stability and power efficiency as well. I would recommend this for your case.

However, by comparing raw performance, the i9 7980XE is undoubtedly the faster CPU, with more cores, a...


OK, but what is the difference in floating point benchmarks between the two chips? Has anyone here run Geekbench 64 to compare these two processors? I'm looking for the fastest floating point CPU I can afford.
 
will depend on your specific application, 12 cores at higher clock vs. 18 cores at lower clock speed...

You might also want to throw in the 16 core $1000 Threadripper 1950X for value comparison.....

As the Xeon Gold is $3100 and up, I'd research your intended applications' scaling with cores vs. clockspeed quite carefully...especially if you are the one paying for it...

Sometimes TR4 is quite ideal....!
 
https://www.servethehome.com/amd-epyc-7501-benchmarks-and-review-32-cores-per-socket/

And since you have a $3000+ cpu budget, one certainly would wnat to consider the 24/32 core Epyc processors as well....(typically providing 50% more cores for the same cost as Xeon Gold)
 
Thanks, everyone, for responding. But I'm still in a quandary. I do very high end computations for science, engineering, and management. My current workstation has a Gigabyte GA-X58-UD7 motherboard with an i7-990x CPU overclocked to 4.1 GHz. It was the fastest system I could buy six years ago, but now it's too slow. Prior to that system I had a Supermicro Dual Xeon motherboard--which performed quite well for the time. This is why I'm currently looking to build a modern Dual Xeon System with the ASUS WS Sage motherboard. But, you're right the Xeons are very, very expensive. So I thought I'd look at the X299 motherboards with the i9-7980xe CPU; unfortunately the ASUS X299 motherboards get poor reviews on Amazon. But I cannot find a floating point comparison anywhere.
 
The Xeon Gold 6146 has 2 AVX 512 units per core, which makes them faster for floating point operations that utilise Intel's Advanced Vector eXtensions.

However, putting that aside, since you use it for highly critical business tasks, reliability is key. The i9 7980XE may tend to overload current X299 boards, and may have some very slight instability issues, which are unacceptable in your use case.

In this case, Xeon 6146 is for you as it supports ECC RAM (error correction ram), more RAM (up to 1TB vs 128GB) and is also binned for 24/7 usage, stability and power efficiency as well. I would recommend this for your case.

However, by comparing raw performance, the i9 7980XE is undoubtedly the faster CPU, with more cores, a higher turbo clock, and superior overclockability.
 
Solution