Tuesday, September 17th 2024

AMD AGESA 1.2.0.2 Update Fixes Ryzen 9000 Series Inter-Core Latency Issues

According to new latest testing, the latest AGESA (AMD Generic Encapsulated Software Architecture) update, version 1.2.0.2, promises a significant boost in performance for AMD Ryzen 9000 "Zen 5" processors. This update is targeting one of the most crucial aspects of multi-core processing: inter-core latency. The AGESA 1.2.0.2 update addresses challenges initially reported in AMD's Zen 5 architecture, particularly in scenarios demanding rapid communication between multiple cores. Early reports suggest a remarkable reduction in inter-core latency by up to 58%. According to Overclock.net testing, older AGESA 1.2.0.1A showed the cross-CCD latency at around 180 ns. However, with the new AGESA 1.2.0.2 BIOS, the latency is seemingly around 75 ns.

Interestingly, the update has arrived on the ASUS ROG Crosshair X670E motherboard, with BIOS version 2401. BIOS updates with the latest AGESA 1.2.0.2 are still rolling out, so it will be interesting to see further testing and possible improvements. It could be that the cross-CCD latency has just been reported badly, so final testing will conclude the latency increase from Zen 4 to Zen 5 debate.
Sources: Overclock.net, via VideoCardz
Add your own comment

46 Comments on AMD AGESA 1.2.0.2 Update Fixes Ryzen 9000 Series Inter-Core Latency Issues

#27
Tomorrow
csendesmarkWOW, nice! :D
I guess this fix is improving those numbers too.
The fix is for Zen 5 based dual-CCD Ryzen models. Currents Threadrippers are using Zen 4 and Zen 5 based Threadrippers are not yet released.
The 80-90ns for Zen 4 based Threadrippes is already normal and in line with Zen 4 based dual-CCD Ryzen models.
Posted on Reply
#28
tfp
TomorrowIt will not improve gaming.

It does not. As you can see from the images inter-core latencies are the same and besides it's hard to improve below 20ns anyway.

www.anandtech.com/show/21124/amd-ryzen-threadripper-7980x-and-7970x-review/4
It could improve gaming on 12 and 16 core CPUs if windows is splitting threads across CCDs but it's only going to help performance in gaming vs the 79XX chips.
Posted on Reply
#29
Tomorrow
tfpIt could improve gaming on 12 and 16 core CPUs if windows is splitting threads across CCDs but it's only going to help performance in gaming vs the 79XX chips.
In the case you describe there's already performance loss vs normal and it would be better fixed by proper thread scheduling via gamebar/chipset driver like AMD recommends.
Posted on Reply
#30
rusTORK
Finally i found this software. Made heat map for AMD Ryzen 7 5800X3D just for fun.
Posted on Reply
#31
phints
AMD really botched the Zen 5 release. Let's see the process... release 9700X that gets smoked by the 7800X3D. Update Chipset drivers- see performance gain. Wait for AGESA update allowing 105W TDP - performance gain. Wait for Windows 11 24H2 - performance gain. Wait for second AGESA update improving latency - performance gain.

At some point 9700X will have all this stuff out and drop in price a bit and be far more interesting, maybe by end of year.
Posted on Reply
#32
windwhirl
phintsLet's see the process... release 9700X that gets smoked by the 7800X3D.
Because the people looking at the 9700X are not supposed to be gamers and are not supposed to care for the V-cache.

It's a case of people looking at the wrong product for their needs.
phintsWait for AGESA update allowing 105W TDP - performance gain.
Little performance gain for like nearly double TDP is so dumb.
Posted on Reply
#33
ajcardiac
Darc RequiemAMD should have just pushed back Zen 5's launch a couple of months and ironed out all these issues. Intel is having issues right now, so I don't see the impetus behind the rush to market.
Indeed and they could have launched it alongside the new chipset motherboards
Posted on Reply
#34
Chry
Sorry but:
Intel: Releases suicidal CPUs only to nerf them after all the reviews
AMD: Releases handicapped CPUs only to boost them after all the reviews

In the end, I would REALLY like to see some proper NEW benchmarks of intel gen 14 and amd zen5 compared side by side both with the LATEST bios and windows versions tested.
If anyone know of any such benchmarking videos/pages please reply to me here or send a PM, much appreciated!
Posted on Reply
#36
dgianstefani
TPU Proofreader
ChrySorry but:
Intel: Releases suicidal CPUs only to nerf them after all the reviews
AMD: Releases handicapped CPUs only to boost them after all the reviews

In the end, I would REALLY like to see some proper NEW benchmarks of intel gen 14 and amd zen5 compared side by side both with the LATEST bios and windows versions tested.
If anyone know of any such benchmarking videos/pages please reply to me here or send a PM, much appreciated!
TPU tests at manually set Intel TDP numbers not mobo makers hallucinated 4096 W limits/no turbo boost limit etc. So from day 1 TPU numbers are/were accurate.

Regardless, even with mobo "stock" settings there was very little change in gaming performance with the enforced limits in new BIOS.
Posted on Reply
#37
Hecate91
TPU testing numbers aren't accurate if they aren't tested at stock, stock is whatever out of the box settings the mobo is using, as most people won't be messing around in the bios to change wattage limits. Intel allowing board makers to set wattage limits too high is partly how they got in this mess.
And Intel claimed there was no performance impact, yet the impact is up to 10%, though it isn't so much about the performance impact, IMO, its more of the questionable reliability of any of the Intel 13th or 14th gen cpu's as Intel tried to hide the issues for over a year.
Posted on Reply
#38
Minus Infinity
OnasiNot necessarily, it says inter-core, not solely cross-CCD. So it should help single CCD chips too, albeit to a lesser extent, theoretically.
Inter-core literally means cores from different ccd's otherwise it would read intra-core. This will have no impact on Ryzen 5 or 7.
Posted on Reply
#39
redzo
Minus InfinityInter-core literally means cores from different ccd's otherwise it would read intra-core. This will have no impact on Ryzen 5 or 7.
Hm. Isn't intra-core supossed to read inside the core ?
Inter CCD - cores between CCD's
Intra CCD - cores inside CCD
Inter core - cores between any cores, from same CCD or between CCD's
:roll: Or do I need my morning cofee ?

I think the gaming workload was already going on a single CCD for 9900x/9950x, same as 7900x3d/7950x3D. This means no improvement in gaming workloads also for 9900x/9950x.
Maybe small gains are expected in some highly threaded cherry picked scenarios as others pointed out, but only for for the dual CCD parts.
Posted on Reply
#40
tfdsaf
Beginner Macro DeviceDoes this mean they will become better in gaming?
I'm no expert, this is a serious question. I wanna know.
Most likely there will be improvement in certain games and certain scenarios where that is indeed the bottleneck. I do expect to see a minor performance bump in many games, but there are likely to be a smaller, but still significant number of games where the performance is going to be improved quite a bit.

I expect overall in an average in 40+ games the CPU's to gain around 3% better performance due to lower core latencies. Every bit helps though and with the Win11 update adding around 4-5% more performance, this probably adding 2-3% more performance, overall that ends up at around 6-8% overall more performance than when Zen5 cpu's launched.
Posted on Reply
#41
Launcestonian
AssimilatorAnd there is this thing called "extensive quality assurance testing" that you do, as a responsible company, to ensure your users don't experience those pains. Unfortunately every company is making customers their beta testers nowadays.
More like a side effect from capitalism. Zen 5 launch timing designed to hit the competition where it hurts cause' when they are bleeding from the Raptor Lake problems, you know the market wants something new & better even if its not perfect at the time.
Posted on Reply
#42
phints
tfdsafMost likely there will be improvement in certain games and certain scenarios where that is indeed the bottleneck. I do expect to see a minor performance bump in many games, but there are likely to be a smaller, but still significant number of games where the performance is going to be improved quite a bit.

I expect overall in an average in 40+ games the CPU's to gain around 3% better performance due to lower core latencies. Every bit helps though and with the Win11 update adding around 4-5% more performance, this probably adding 2-3% more performance, overall that ends up at around 6-8% overall more performance than when Zen5 cpu's launched.
Agree with this, gains will be small and isolated to certain titles. And likely only 9600X and 9900X will benefit since they are multi-CCD CPUs. 9700X should have little to no benefit from what I can tell. Ryzen has used 8C/16T per CCD for 3 generations.
Posted on Reply
#43
mechtech
"Early reports suggest a remarkable reduction in inter-core latency by up to 58%."

But can it do that to my 5700x cpu?? :)
Posted on Reply
#44
bikemanI7
Any Benefit to Ryzen 7700X Chip? and does this AGESA 1.2.0.2 include the Sinkhole Vulnerability fix, My MSI Pro B650 VC Wifi just went final this morning, now to find the day without Real Life Interruptions and get it installed asap
Posted on Reply
#45
windwhirl
bikemanI7Any Benefit to Ryzen 7700X Chip?
I suppose not. The improvement is mostly for "out of CCD" communications.
bikemanI7and does this AGESA 1.2.0.2 include the Sinkhole Vulnerability fix,
I think that was fixed in 1.2.0.1
Posted on Reply
#46
bikemanI7
Yeah well My Motherboard Manufacturer didn't have the UEFI Bios with AGESA 1.2.0.1 one out of beta, they do have version with AGESA 1.2.0.2 out of beta though as of this morning, so eventually will get the 1.2.0.2 version installed here and all should be good hopefully
Posted on Reply
Add your own comment
Dec 4th, 2024 03:33 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts