Timeline for NSIGHT compute: SOL SM versus Roofline
Current License: CC BY-SA 4.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 10, 2021 at 10:59 | history | edited | einpoklum |
edited tags
|
|
| Jan 9, 2021 at 6:28 | history | edited | Bram | CC BY-SA 4.0 |
Update with likely cause.
|
| Jan 9, 2021 at 5:15 | answer | added | Robert Crovella | timeline score: 4 | |
| Jan 9, 2021 at 2:56 | history | edited | Bram | CC BY-SA 4.0 |
Add peak stat.
|
| Jan 9, 2021 at 2:43 | comment | added | Bram | @talonmies Ah, thank you. Indeed: I read the values wrong, off by a factor 1000. It was GFlops, not TFlops. Still my observation stands: The SoL says: close to max. The RoofLine says: not even remotely close to max. | |
| Jan 9, 2021 at 2:42 | history | edited | Bram | CC BY-SA 4.0 |
Fix postfix: gigaflops, not teraflops.
|
| Jan 9, 2021 at 1:45 | comment | added | talonmies | I think you need to check the magnitude of the numbers in your question because they don't look even remotely realistic | |
| Jan 8, 2021 at 17:04 | history | asked | Bram | CC BY-SA 4.0 |