pututu
[H]ard DC'er of the Year 2021
- Joined
- Dec 27, 2015
- Messages
- 3,157
Phoenicis, thanks for un-hiding your threadripper stats in mindmodeling project. Seen a lot of synthetic benchmarking numbers but would like to share how TR performs in BOINC projects in our private forum first.
I'm comparing stats from your rig here and mine is here from Windows only tasks for MindModeling.
pututu Xeon 2683 v3, Win 7, 14C/28T @100%, 3GHz all cores, DDR4 at 2133MHz.
phoenics, Ryzen Threadripper 1950X, Win 10, 16C/32T, 3.7GHz(?)
I took the valid task results from most recent 120 tasks. This should be statistically significant, I hope. Below is what I found. Note that the numbers in the table below represents credits per hour for a single task. I'm comparing core to core performance.
TR floating point speed and integer speed (again synthetic benchmark numbers) blow away my 2683v3 but in MindModeling project, the gain is much less, around +13%. I'm not sure if this is due to xeon more efficient instruction sets, different BOINC version used (7.6.33 vs 7.7.2), Windows version, etc.
Avg power draw from the outlet for my xeon is currently 191W.
I'm comparing stats from your rig here and mine is here from Windows only tasks for MindModeling.
pututu Xeon 2683 v3, Win 7, 14C/28T @100%, 3GHz all cores, DDR4 at 2133MHz.
phoenics, Ryzen Threadripper 1950X, Win 10, 16C/32T, 3.7GHz(?)
I took the valid task results from most recent 120 tasks. This should be statistically significant, I hope. Below is what I found. Note that the numbers in the table below represents credits per hour for a single task. I'm comparing core to core performance.
TR floating point speed and integer speed (again synthetic benchmark numbers) blow away my 2683v3 but in MindModeling project, the gain is much less, around +13%. I'm not sure if this is due to xeon more efficient instruction sets, different BOINC version used (7.6.33 vs 7.7.2), Windows version, etc.
Avg power draw from the outlet for my xeon is currently 191W.