Improvements to the Cache Hierarchy

The biggest under-the-hood change for the Ryzen 2000-series processors is in the cache latency. AMD is claiming that they were able to knock one-cycle from L1 and L2 caches, several cycles from L3, and better DRAM performance. Because pure core IPC is intimately intertwined with the caches (the size, the latency, the bandwidth), these new numbers are leading AMD to claim that these new processors can offer a +3% IPC gain over the previous generation.

The numbers AMD gives are:

  • 13% Better L1 Latency (1.10ns vs 0.95ns)
  • 34% Better L2 Latency (4.6ns vs 3.0ns)
  • 16% Better L3 Latency (11.0ns vs 9.2ns)
  • 11% Better Memory Latency (74ns vs 66ns at DDR4-3200)
  • Increased DRAM Frequency Support (DDR4-2666 vs DDR4-2933)

It is interesting that in the official slide deck AMD quotes latency measured as time, although in private conversations in our briefing it was discussed in terms of clock cycles. Ultimately latency measured as time can take advantage of other internal enhancements; however a pure engineer prefers to discuss clock cycles.

Naturally we went ahead to test the two aspects of this equation: are the cache metrics actually lower, and do we get an IPC uplift?

Cache Me Ousside, How Bow Dah?

For our testing, we use a memory latency checker over the stride range of the cache hierarchy of a single core. For this test we used the following:

  • Ryzen 7 2700X (Zen+)
  • Ryzen 5 2400G (Zen APU)
  • Ryzen 7 1800X (Zen)
  • Intel Core i7-8700K (Coffee Lake)
  • Intel Core i7-7700K (Kaby Lake)

The most obvious comparison is between the AMD processors. Here we have the Ryzen 7 1800X from the initial launch, the Ryzen 5 2400G APU that pairs Zen cores with Vega graphics, and the new Ryzen 7 2700X processor.

This graph is logarithmic in both axes.

This graph shows that in every phase of the cache design, the newest Ryzen 7 2700X requires fewer core clocks. The biggest difference is on the L2 cache latency, but L3 has a sizeable gain as well. The reason that the L2 gain is so large, especially between the 1800X and 2700X, is an interesting story.

When AMD first launched the Ryzen 7 1800X, the L2 latency was tested and listed at 17 clocks. This was a little high – it turns out that the engineers had intended for the L2 latency to be 12 clocks initially, but run out of time to tune the firmware and layout before sending the design off to be manufactured, leaving 17 cycles as the best compromise based on what the design was capable of and did not cause issues. With Threadripper and the Ryzen APUs, AMD tweaked the design enough to hit an L2 latency of 12 cycles, which was not specifically promoted at the time despite the benefits it provides. Now with the Ryzen 2000-series, AMD has reduced it down further to 11 cycles. We were told that this was due to both the new manufacturing process but also additional tweaks made to ensure signal coherency. In our testing, we actually saw an average L2 latency of 10.4 cycles, down from 16.9 cycles in on the Ryzen 7 1800X.

The L3 difference is a little unexpected: AMD stated a 16% better latency: 11.0 ns to 9.2 ns. We saw a change from 10.7 ns to 8.1 ns, which was a drop from 39 cycles to 30 cycles.

Of course, we could not go without comparing AMD to Intel. This is where it got very interesting. Now the cache configurations between the Ryzen 7 2700X and Core i7-8700K are different:

CPU Cache uArch Comparison
  AMD
Zen (Ryzen 1000)
Zen+ (Ryzen 2000)
Intel
Kaby Lake (Core 7000)
Coffee Lake (Core 8000)
L1-I Size 64 KB/core 32 KB/core
L1-I Assoc 4-way 8-way
L1-D Size 32 KB/core 32 KB/core
L1-D Assoc 8-way 8-way
L2 Size 512 KB/core 256 KB/core
L2 Assoc 8-way 4-way
L3 Size 8 MB/CCX
(2 MB/core)
2 MB/core
L3 Assoc 16-way 16-way
L3 Type Victim Write-back

AMD has a larger L2 cache, however the AMD L3 cache is a non-inclusive victim cache, which means it cannot be pre-fetched into unlike the Intel L3 cache.

This was an unexpected result, but we can see clearly that AMD has a latency timing advantage across the L2 and L3 caches. There is a sizable difference in DRAM, however the core performance metrics are here in the lower caches.

We can expand this out to include the three AMD chips, as well as Intel’s Coffee Lake and Kaby Lake cores.

This is a graph using cycles rather than timing latency: Intel has a small L1 advantage, however the larger L2 caches in AMD’s Zen designs mean that Intel has to hit the higher latency L3 earlier. Intel makes quick work of DRAM cycle latency however.

Talking 12nm and Zen+ Translating to IPC: All This for 3%?
POST A COMMENT

544 Comments

View All Comments

  • jjj - Thursday, April 19, 2018 - link

    I was wondering about gaming, so there is no mistake there as Ryzen 2 seems to top Intel.
    As of right now, I don't seem to find memory specs in the review yet, safe to assume you did as always, highest non-OC so Ryzen is using faster DRAM?
    Also yet to spot memory letency, any chance you have some numbers at 3600MHz vs Intel? Thanks.
    Reply
  • jjj - Thursday, April 19, 2018 - link

    And just between us, would be nice to have some Vega gaming results under DX12. Reply
  • aliquis - Thursday, April 19, 2018 - link

    Would be nice if any reviewer actually benchmarked storage devices maybe even virtualization because then we'd see meltdown and spectre mitigation performance. Then again do AMD have any for spectre v2 yet? If not who knows what that will do. Reply
  • HStewart - Thursday, April 19, 2018 - link

    I notice that that systems had higher memory, but for me I believe single threaded performance is more important that more cores. But it would be bias if one platform is OC more than another. Personally I don't over clock - except for what is provided with CPU like Turbo mode.

    One thing that I foresee in the future is Intel coming out with 8 core Coffee Lake

    But at least it appears one thing is over is this Meltdown/Spectre stuff
    Reply
  • Lolimaster - Thursday, April 19, 2018 - link

    Intel 8 core CL won't stop the bleeding, lose more profits making them "cheap" vs a new Ryzen 7nm with at least 10% more clocks and 10% more IPC, RIP. Reply
  • HStewart - Thursday, April 19, 2018 - link

    I just have to agree to disagree on that statement - especially on "cheap" statement Reply
  • ACE76 - Thursday, April 19, 2018 - link

    CL can't scale to 8 cores...not without done serious changes to it's architecture...Intel is in some trouble with this Ryzen refresh...also worth noting is that 7nm Ryzen 2 will likely bring a considerable performance jump while Intel isn't sitting on anything worthwhile at the moment. Reply
  • Alphasoldier - Friday, April 20, 2018 - link

    All Intel's 8cores in HEDT except SkylakeX are based on their year older architecture with a bigger cache and the quad channel.

    So if Intel have the need, they will simply make a CL 8core. 2700X is pretty hungry when OC'd, so Intel don't have to worry at all about its power consuption.
    Reply
  • moozooh - Sunday, April 22, 2018 - link

    > 2700X is pretty hungry when OC'd
    And Intel chips aren't? If Zen+ is already on Intel's heels for both performance per watt and raw frequency, a 7nm chip with improved IPC and/or cache is very likely going to have them pull ahead by a significant margin. And even if it won't, it's still going to eat into Intel's profit as their next tech is 10nm vs. AMD's 7nm, meaning more optimal wafer estate utilization for the latter.

    AMD has really climbed back at the top of their game; I've been in the Intel camp for the last 12 years or so, but the recent developments throw me way back to K7 and A64 days. Almost makes me sad that I won't have any reason to move to a different mobo in the next 6–8 years or so.
    Reply
  • mapesdhs - Friday, March 29, 2019 - link

    Amusing to look back given how things panned out. So yes, Intel released the 9900K, but it was 100% more expensive than the 2700X. :D A complete joke. And meanwhile tech reviewers raved about a peasly 5 to 5.2 oc, on a chip that already has a 4.7 max turbo (major yawn fest), focusing on specific 1080p gaming tests that gave silly high fps number favoured by a market segment that is a tiny minority. Then what happens, RTX comes out and pushes the PR focus right back down to 60Hz. :D

    I wish people to stop drinking the Intel/NVIDIA coolaid. AMD does it aswell sometimes, but it's bizarre how uncritical tech reviewers often are about these things. The 9900K dragged mainstream CPU pricing up to HEDT levels; epic fail. Some said oh but it's great for poorly optimised apps like Premiere, completely ignoring the "poorly optimised" part (ie. why the lack of pressure to make Adobe write better code? It's weird to justify an overpriced CPU on the back of a pro app that ought to run a lot better on far cheaper products).
    Reply

Log in

Don't have an account? Sign up now