By: Foyle (nope.delete@this.nope.com), September 19, 2022 4:08 am
Room: Moderated Discussions
anonymous2 (anonymous2.delete@this.example.com) on September 18, 2022 2:40 pm wrote:
> https://github.com/nviennot/core-to-core-latency
>
> Why is Intel Core i9-9900K so low vs other Intel cores?
>
Table at the top of the page has wrong data for Skylake (i7-6700K). Graph below seems correct. There are no significant differences between *lake generations if you normalize the ring frequency (main variable, but not accounted for in the linked repo) and if you take the ring length (counted by # of stops) into account. In fact with 8-core Coffee Lake and 10-core Comet Lake the median core-to-core latency got worse.
> https://github.com/nviennot/core-to-core-latency
>
> Why is Intel Core i9-9900K so low vs other Intel cores?
>
Table at the top of the page has wrong data for Skylake (i7-6700K). Graph below seems correct. There are no significant differences between *lake generations if you normalize the ring frequency (main variable, but not accounted for in the linked repo) and if you take the ring length (counted by # of stops) into account. In fact with 8-core Coffee Lake and 10-core Comet Lake the median core-to-core latency got worse.
| Topic | Posted By | Date |
|---|---|---|
| GH repo to measure core-core latency | anonymous2 | 2022/09/18 02:40 PM |
| GH repo to measure core-core latency | Foyle | 2022/09/19 04:08 AM |
| GH repo to measure core-core latency | Michael S | 2022/09/20 02:56 AM |
| GH repo to measure core-core latency | Brendan | 2022/09/19 06:14 AM |
| GH repo to measure core-core latency | anon2 | 2022/09/19 08:14 PM |