Article Index
Time has finally come.
Today we are bringing you our full review of the Threadripper 2990WX and Threadripper 2950X.
So they are based on the Zen+ architecture which uses the 12PL process from GlobalFoundries.
When using DDR4-3200 memory the Infinity Fabric throughput between these dies is roughly 50 GBps.
As was the case with the 1950X, the 2950X can be configured in one of two ways.
The 2990WX on the other hand is a very different beast.
It consists of not two Zeppelin dies but rather four, enabling up to 32 cores.
However on the X399 platform AMD has imposed some limitations to avoid cannibalizing their single socketEPYC server CPUs.
The biggest of these limitations being that there are still just four memory controllers.
The original Threadripper chips didn’t have ‘dummy’ dies as claimed by AMD.
The problem is memory bandwidth, as there simply isn’t going to be enough of it.
Basically all systems were configured with 32GB of DDR4-3200 memory using XMP timings.
So let’s get into the results.
Anyway in this particular synthetic rendering benchmark the 2990WX has no trouble blowing socks off.
Next up we have another rendering benchmark though this one is based on real-world software.
The Corona Renderer has been used to test workstations with over 64 cores, so it scales very well.
This also meant it was 28% faster than Intel current flagship Core i9 part.
So what about a workload that takes significantly longer than a few seconds for the 2990WX to complete.
Disappointingly the much more complex and therefore longer to complete Gooseberry workload was less favorable to the 2990X.
It was also 57% faster than the Core i9-7980XE, so a massive win there.
Here we see the 2990WX providing a surprisingly poor result taking 43 seconds to complete the workload.
This made the 32-core processor slower than even the 1950X.
Here it was the 2950X that impressed, matching the Core i9-7960X and 7980XE.
