What is this page?
In this page we encourage all GPUGRUD users to test their GPU performance by comparing the completion time of their work units (WU) with other users. WU are splitted into batches, as different simulation setups may yield variance in computation length and graphic card performance. Only WU completed in the last 15 days and batches with more than 50 completed units are taken into account in the analysis.
The box plot expresses the variance of data per batch. Find out its maximum, minimum, average time and the number of samples (WU) in the group by hovering over it.
If you are logged in and have recently crunched workunits, your personal records will be calculated. The quickest WU per task and your position expressed in quartiles (from faster 1 to slower 4) are shown.
You can also browse the top thirty timestamps for each batch.
NEW: We have included a new ranking that shows the best average performers in the last week with a minimum amount of 10 WU crunched.
Top average performers (last week Long Runs)
|Rank||User name||Average (h)||Total crunched (WU)||GPU description of fastest setup|
|1||CoolAtchOk||4.67000||20|| NVIDIA GeForce GTX 1080 Ti (4095MB) driver: 399.24|
|2||Retvari Zoltan||5.02529||87||NVIDIA GeForce GTX 1080 Ti (4095MB) driver: 436.48|
|3||compalex||8.42143||42||NVIDIA GeForce GTX 1080 (4095MB) driver: 417.35|
|4||Rion Family||10.53333||18|| NVIDIA GeForce GTX 1080 (4095MB) driver: 430.86|
|5||RKN-Cluster||13.00000||21|| NVIDIA GeForce GTX 1060 3GB (3072MB) driver: 436.30|
|6||Erich56||16.41739||23|| NVIDIA GeForce GTX 980 Ti (4095MB) driver: 430.86|
Personal records (Log in)
|WU batch name||Best time (h)||Quartile (1-4)|
Top performers per batch
DISCLAIMER: inconsistencies in statistics and rankings may be due to privacy issues. Be sure to enable the "view computers" option in your profile if you want to be ranked. Overclock responsibly.