By: hobold (hobold.delete@this.vectorizer.org), October 10, 2020 12:19 pm
Room: Moderated Discussions
Maynard Handley (name99.delete@this.name99.org) on October 10, 2020 10:36 am wrote:
> hobold (hobold.delete@this.vectorizer.org) on October 10, 2020 8:58 am wrote:
[...]
> > But one cannot reduce latency below zero (no matter what marketing departments
> > say), so these gains are not very likely to be repeated by AMD.
>
> Yes you can :-) Value prediction is real, and done competently provides genuine speedups...
> As far as I can tell no-one is doing value prediction yet (as always asterisk against Apple...)
> but it's still a promising technology once everything easier has been done.
Latency is the time you must wait for an unknown result that will be produced in the future. Value prediction is re-use of a known result that has been produced in the past.
It is possible to try and guess the future. It is not commonly possible to reliably violate causality (in the macroscopic world).
I do not mean to diminish all the ingenious speculative algorithms. But for the benefit of the lurking readers, I want to re-iterate that latency times are always strictly greater than zero.
No matter what marketing departments or internet discussion forums say. :-)
> hobold (hobold.delete@this.vectorizer.org) on October 10, 2020 8:58 am wrote:
[...]
> > But one cannot reduce latency below zero (no matter what marketing departments
> > say), so these gains are not very likely to be repeated by AMD.
>
> Yes you can :-) Value prediction is real, and done competently provides genuine speedups...
> As far as I can tell no-one is doing value prediction yet (as always asterisk against Apple...)
> but it's still a promising technology once everything easier has been done.
Latency is the time you must wait for an unknown result that will be produced in the future. Value prediction is re-use of a known result that has been produced in the past.
It is possible to try and guess the future. It is not commonly possible to reliably violate causality (in the macroscopic world).
I do not mean to diminish all the ingenious speculative algorithms. But for the benefit of the lurking readers, I want to re-iterate that latency times are always strictly greater than zero.
No matter what marketing departments or internet discussion forums say. :-)
Topic | Posted By | Date |
---|---|---|
Zen 3 | Blue | 2020/10/08 09:58 AM |
Zen 3 | Rayla | 2020/10/08 10:10 AM |
Zen 3 | Adrian | 2020/10/08 10:13 AM |
Does anyone know whether Zen 3 has AVX-512? (NT) | Foo_ | 2020/10/08 11:54 AM |
Does anyone know whether Zen 3 has AVX-512? | Adrian | 2020/10/08 12:11 PM |
Zen 3 - Number of load/store units | ⚛ | 2020/10/08 10:21 AM |
Zen 3 - Number of load/store units | Rayla | 2020/10/08 10:28 AM |
Zen 3 - Number of load/store units | ⚛ | 2020/10/08 11:22 AM |
Zen 3 - Number of load/store units | Adrian | 2020/10/08 11:53 AM |
Zen 3 - Number of load/store units | Travis Downs | 2020/10/08 09:45 PM |
Zen 3 - CAD benchmark | Per Hesselgren | 2020/10/09 07:29 AM |
Zen 3 - CAD benchmark | Adrian | 2020/10/09 09:27 AM |
Zen 3 - Number of load/store units | itsmydamnation | 2020/10/08 02:38 PM |
Zen 3 - Number of load/store units | Groo | 2020/10/08 02:48 PM |
Zen 3 - Number of load/store units | Wilco | 2020/10/08 03:02 PM |
Zen 3 - Number of load/store units | Dummond D. Slow | 2020/10/08 04:39 PM |
Zen 3 - Number of load/store units | Doug S | 2020/10/09 08:11 AM |
Zen 3 - Number of load/store units | Dummond D. Slow | 2020/10/09 09:43 AM |
Zen 3 - Number of load/store units | Doug S | 2020/10/09 01:43 PM |
N7 and N7P are not load/Store units - please fix the topic in your replies (NT) | Heikki Kultala | 2020/10/10 07:37 AM |
Zen 3 | Jeff S. | 2020/10/08 12:16 PM |
Zen 3 | anon | 2020/10/08 01:57 PM |
Disappointing opening line in paper | Paul A. Clayton | 2020/10/11 06:16 AM |
Thoughts on "Improving the Utilization of µop Caches..." | Paul A. Clayton | 2020/10/14 12:11 PM |
Thoughts on "Improving the Utilization of µop Caches..." | anon | 2020/10/15 11:56 AM |
Thoughts on "Improving the Utilization of µop Caches..." | anon | 2020/10/15 11:57 AM |
Sorry about the mess | anon | 2020/10/15 11:58 AM |
Sorry about the mess | Brett | 2020/10/16 03:22 AM |
Caching dependence info in µop cache | Paul A. Clayton | 2020/10/16 06:20 AM |
Caching dependence info in µop cache | anon | 2020/10/16 12:36 PM |
Caching dependence info in µop cache | Paul A. Clayton | 2020/10/18 01:28 PM |
Zen 3 | juanrga | 2020/10/09 10:12 AM |
Zen 3 | Mr. Camel | 2020/10/09 06:30 PM |
Zen 3 | anon.1 | 2020/10/10 12:44 AM |
Cinebench is terrible benchmark | David Kanter | 2020/10/10 10:36 AM |
Cinebench is terrible benchmark | anon.1 | 2020/10/10 12:06 PM |
Cinebench is terrible benchmark | hobold | 2020/10/10 12:33 PM |
Some comments on benchmarks | Paul A. Clayton | 2020/10/14 12:11 PM |
Some comments on benchmarks | Mark Roulo | 2020/10/14 03:21 PM |
Zen 3 | Adrian | 2020/10/10 01:59 AM |
Zen 3 | Adrian | 2020/10/10 02:18 AM |
Zen 3 | majord | 2020/10/15 04:02 AM |
Zen 3 | hobold | 2020/10/10 08:58 AM |
Zen 3 | Maynard Handley | 2020/10/10 10:36 AM |
Zen 3 | hobold | 2020/10/10 12:19 PM |
Zen 3 | anon | 2020/10/11 02:58 AM |
Zen 3 | hobold | 2020/10/11 12:32 PM |
Zen 3 | anon | 2020/10/11 01:07 PM |
Zen 3 | hobold | 2020/10/11 02:22 PM |
Zen 3 | anon | 2020/10/10 11:51 AM |
Zen 3 | Michael S | 2020/10/11 01:16 AM |
Zen 3 | hobold | 2020/10/11 02:13 AM |
Zen 3 | Michael S | 2020/10/11 02:18 AM |
Zen 3 | anon.1 | 2020/10/11 12:17 PM |
Zen 3 | David Hess | 2020/10/12 06:43 AM |
more power? (NT) | anonymous2 | 2020/10/12 01:26 PM |
I think he's comparing 65W 3700X vs 105W 5800X (NT) | John H | 2020/10/12 04:33 PM |
?! Those are apples and oranges! (NT) | anon | 2020/10/12 04:49 PM |