We’ve got some newer slaves that are not performing to our expectations. The monitor is not reporting the full number of cores that we expect to be seeing. When we submit a job we see all logical cores engage but the performance between that node and an older node we are comparing to is essentially the same.
3d116: legacy system running well
Windows 7 Version 6.1.7601 Service Pack 1 Build 7601
Processor Intel® Xeon® CPU E5-2690 v3 @ 2.60GHz, 2601 Mhz, 12 Core(s), 24 Logical Processor(s)
Processor Intel® Xeon® CPU E5-2690 v3 @ 2.60GHz, 2601 Mhz, 12 Core(s), 24 Logical Processor(s)
Installed Physical Memory (RAM) 128 GB
Page File Space 128 GB
Seagate 2 TB, SATA, 7200 RPM drive
3d201: render times using all cores greater than or equal to 3d101
Windows 10 Version 10.0.16299 Build 16299
Processor Intel® Xeon® Gold 6148 CPU @ 2.40GHz, 2401 Mhz, 20 Core(s), 40 Logical Processor(s)
Processor Intel® Xeon® Gold 6148 CPU @ 2.40GHz, 2401 Mhz, 20 Core(s), 40 Logical Processor(s)
Installed Physical Memory (RAM) 384 GB
Page File Space 40.0 GB
Samsung PM863a, 960GB,SATA SSD
We are currently running Deadline 9, I have installed 10 on a test node and see the same results. We are using 3ds Max and Vray. I’ve run Vray bench mark and see baseline of 25 seconds on the newer 200 node compared to 40 seconds on the older 100 node and all cores are detected.
We are seeing comparable performance with and without cpu affinity on.
Quick note, I’ve referred to 3d101 and 116 both in this post, nodes 101-116 are all the same hardware build and perform the same way.