←back to thread

Evaluating the Infinity Cache in AMD Strix Halo

(chipsandcheese.com)
141 points zdw | 1 comments | 22 Oct 25 04:20 UTC | HN request time: 0s | source
Show context
joelthelion ◴[22 Oct 25 06:15 UTC] No.45665436[source]▶
I don't quite get it. What's so special about having 32MB of cache? Why is it called "infinity"?
replies(3): >>45665469 #>>45665483 #>>45665667 #
phire ◴[22 Oct 25 06:50 UTC] No.45665667[source]▶
AMD named their memory fabric "infinity fabric" for marketing reasons. So when they developed their memory attached cache solution (which lives in the memory fabric, unlike a traditional cache), the obvious marketing name is "infinity cache"

The main advantage of a memory attached cache is that it's cheaper than a regular cache, and can even be put on a seperate die, allowing you to have much more of it.

AMDs previous memory fabric from the early 2000s was called "Hyper Transport", which has a confusing overlap with Intel's Hyper Threading, but I think AMD actually bet intel to the name by a few years.

replies(4): >>45665841 #>>45666573 #>>45673803 #>>45676022 #
typpilol ◴[22 Oct 25 07:22 UTC] No.45665841[source]▶
How's latency vs a traditional?
replies(1): >>45666039 #
1. Tuna-Fish ◴[22 Oct 25 07:54 UTC] No.45666039[source]▶
It has a higher latency than AMD's L2 caches, but similar compared to Nvidia's L2. [0]

I think AMD might be ditching it for the next gen in exchange for growing the L2's, because the lower latency is beneficial to ray tracing.

[0]: https://substackcdn.com/image/fetch/$s_!LHJo!,f_auto,q_auto:...