Thursday, August 3rd 2023

AMD Readying AGESA 1.0.0.7c for AM5 Motherboards

According to a post by @g01d3nm4ng0 on Twitter/X, we now know that AMD is readying yet another AGESA update for AM5 motherboards. The new version is, based on information from our own sources, a minor update to the current version. As such, AMD will be moving from 1.0.0.7b to 1.0.0.7c. @g01d3nm4ng0 didn't reveal any details of the new AGESA apart from the screenshot below, but we asked around and managed to find out what the new AESA addresses.

The update is specifically for those with Samsung DDR5 memory in their AM5 motherboards and it addresses multiple memory related stability issues. We weren't given the full details as to what those are, but there have been some reports about there being issues specifically with Samsung DDR5 memory in some AM5 boards and hopefully this will solve all those problems. We don't have a release time frame for the updated AGESA, but with 1.0.0.7b barely out the door, it might take a few weeks before this one makes it through all the internal testing at the motherboard makers.
Source: @g01d3nm4ng0 on Twitter/X
Add your own comment

56 Comments on AMD Readying AGESA 1.0.0.7c for AM5 Motherboards

#26
trparky
The problem with being on the bleeding edge...

If I had known that Zen 4 would have so many issues to still iron out when I first bought my parts, I'd have waited a little bit longer as well.
Posted on Reply
#27
A Computer Guy
bugI have no idea why AGESA has such a troubled history. It's supposed to handle just the initialization of a few things.

At one point, Microsoft had to patch Win95 because during init there was a hardcoded waiting time somewhere. As CPUs got faster, other things started up faster and that delay started to get in the way of other things.
I think it's called rush to market then fix it as quickly as possible.
Posted on Reply
#28
trparky
A Computer GuyI think it's called rush to market then fix it as quickly as possible.
They ain't the only ones that do that kind of stuff.
Posted on Reply
#29
A Computer Guy
Minus InfinityThe take home from all this is never buyand AMD system until at least 12 months after MB's have been released. AFter being burned by memory issues with Zen 1, I waited 2 years after Zen 3 shipped to upgrade.
Yes this seems to be good advice but perhaps not even strictly for AMD. Twelve months is a good lead time to wait for issues to crop up and also get a feel for which vendors did a good or poor implementation based on reviews and customer feedback as well as waiting for the bleeding edge bugs to get ironed out. I got in at Zen+ after memory issues were stabilizing and/or at least enough info how to tune the RAM was very helpful to resolve problems especially with Corsair DDR4. I really was excited about the prospect of getting a 7950x but 8950x/9950x will hopefully be worth the wait. If AMD continues the trend of long socket cycles then getting into, or upgrading to, a particular AM socket platform at mid cycle will be a win for the typical consumer that doesn't want to deal with growing pains.
Posted on Reply
#30
TheLostSwede
News Editor
trparkyBut why? What is Samsung doing with their DDR5 memory as versus the other manufacturers? What is Samsung doing wrong? Are they not holding themselves to the DDR5 specifications?
I didn't get the full details, sorry, but it's most likely something AMD missed to address in the previous AGESA release.
Doesn't mean anyone is doing something out of spec, it could simply be something AMD discovered that they had missed or was given some new information that hadn't arrived in time for the previous release.
Keep in mind that although JEDEC sets the specs on how each release of DRAM is supposed to work, it doesn't mean the implementations by each DRAM manufacturer is identical, as it's based on their knowhow, nodes and previous generations of DRAM, so there will be small differences and AMD, Intel etc. has to work around those.
Also, the previous release was about higher memory clocks, outside of JEDEC spec, so this is most likely related to that.
Posted on Reply
#31
A Computer Guy
trparkyThey ain't the only ones that do that kind of stuff.
In my mind I think Microsoft made it popular/normalized way of business but there are probably examples before them that I am not familiar with. It's a tough balance because if you rush it out before it's "ready" a poor product will degrade your brand, too late and your competition might end up taking the market share you could have had.
Posted on Reply
#32
trparky
A Computer GuyIn my mind I think Microsoft
I didn't want to say it but... yeah, it's Microsoft.
Posted on Reply
#33
Tropick
Why_MeAMD and RAM issues go together like coffee and creamer.
It seems to be the curse that just keeps following them around. It really feels like the best time to buy AMD parts is at the end of the generation once they've ironed out all the bugs and have had time to tune everything properly. The AMD early adopter tax has been especially high this gen in both price and bugs.
Posted on Reply
#34
trparky
TropickIt seems to be the curse that just keeps following them around. It really feels like the best time to buy AMD parts is at the end of the generation once they've ironed out all the bugs and have had time to tune everything properly. The AMD early adopter tax has been especially high this gen in both price and bugs.
I agree. AM5 was definitely a bumpy road at first.
Posted on Reply
#35
A Computer Guy
TropickIt seems to be the curse that just keeps following them around. It really feels like the best time to buy AMD parts is at the end of the generation once they've ironed out all the bugs and have had time to tune everything properly. The AMD early adopter tax has been especially high this gen in both price and bugs.
In my opinion I think by Zen2 (the cycle midpoint) really most of the issues got ironed out. Desktop Lifecycle Recap: Zen -> Zen+ -> Lisa Sue Edition -> Zen2 -> Zen3 -> ZenX3D ( -> Zen Overdrive [TBD] )
Posted on Reply
#36
TheLostSwede
News Editor
I know I'm kind of "late" to join the AM5 platform, but I had exactly zero issues getting the RAM I bought to boot at 6000 MHz and that was non QVL modules with XMP, as they don't have EXPO support. I was surprised to be honest, as I've struggled with memory on both my previous AM4 builds. The Ryzen 7 1700 never ended up working properly with the Corsair LPX modules I had, the Ryzen 7 3800X took about six months until the Patriot Viper Steel modules worked at they should, but swapping to the 5800X didn't cause any issues.
Posted on Reply
#37
trparky
I went through two memory kits, both G.Skill, until I got a stable system.

I'm a fan of Hardware Unboxed on YouTube and each of Steve's videos mentions a specific G.Skill kit that he uses in all of his AM5-related benchmark videos. I figured that if he uses that specific kit, I should too. And so, I bought that kit and what do you know? No memory training issues. No having to reset BIOS. Nothing of the sort. Rock-solid stable. I enabled EXPO and away I went.
Posted on Reply
#38
tpa-pr
Interesting, so far so good on the Corsair Vengeance DDR5-6000 kit with XMP turned on. That being said I haven't gone out of my way to stress test it but i'm also not facing any instability. I'll be keeping an eye on it though.
Posted on Reply
#39
Chry

Works flawlessly even OCd to 6000mhz.
Worked flawless with previous BIOS versions since PC first assembled a few months ago, too.
Posted on Reply
#40
Minus Infinity
TheLostSwedeI know I'm kind of "late" to join the AM5 platform, but I had exactly zero issues getting the RAM I bought to boot at 6000 MHz and that was non QVL modules with XMP, as they don't have EXPO support. I was surprised to be honest, as I've struggled with memory on both my previous AM4 builds. The Ryzen 7 1700 never ended up working properly with the Corsair LPX modules I had, the Ryzen 7 3800X took about six months until the Patriot Viper Steel modules worked at they should, but swapping to the 5800X didn't cause any issues.
Weird thing my old X370 board with 1700X had lot's of stability issues even with later bios' (which helped), but since going through the painful process of updating gigabyte bios so it could work with Zen 2/3 CPU's and switching to 3700X, uninstalling all AMD drivers with their cleaner tool, then reinstalling chipset drivers, system has been rock solid now for 6 months.

Whether I upgrade to Zen 5 will depend on not only performance compared to Arrow Lake, but if they have ironed out all the kinks with AM5. I don't mind a little slower if it's rock solid. Hopefully with faster memory being unlocked now for AM5 when Zen 5 comes out with it's rumoured DDR5 6400+ support, memory issues will be a thing of the past.
Posted on Reply
#41
Blados
Cultist
I'm wondering if this revision fixed the power down mode bug introduced with 1007b honestly.
Posted on Reply
#42
Unregistered
BladosI'm wondering if this revision fixed the power down mode bug introduced with 1007b honestly.
what bug? could you please explain
#43
MicroUnC
I got it today

MSI MEG X670E ACE
Description:
- AGESA ComboPI 1.0.0.7c updated. - Mitigate security issue. - Improved DDR5 memory compatibility. - Improved ACPI S3 issue.
Posted on Reply
#44
Blados
Cultist
M440what bug? could you please explain
Disabling Power Down Mode, and enabling Memory Context Restore, completely ruins memory stability after reboot.

as of 1007b, you need to either have both on (and incur the latency penalty that Power Down Mode has), or disable both and suffer excruciating boot times.
Posted on Reply
#45
bug
Minus InfinityThe take home from all this is never buyand AMD system until at least 12 months after MB's have been released. AFter being burned by memory issues with Zen 1, I waited 2 years after Zen 3 shipped to upgrade.
Tbh at the time AMD was upfront about the memory controller being just "good enough" and that incremental improvements will follow. It was one those "don't postpone the whole product/platform launch because of tweaks" kind of decision.

I don't have first hand experience with Zen1, but as far as I understand, it was stable using standard JEDEC profiles. The trouble came that in order to extract the maximum out of Zen, memory needed to be overclocked. That's where trouble started, especially since AMD's XMP was nowhere in sight.
Posted on Reply
#46
Gmr_Chick
trparkyI went through two memory kits, both G.Skill, until I got a stable system.

I'm a fan of Hardware Unboxed on YouTube and each of Steve's videos mentions a specific G.Skill kit that he uses in all of his AM5-related benchmark videos. I figured that if he uses that specific kit, I should too. And so, I bought that kit and what do you know? No memory training issues. No having to reset BIOS. Nothing of the sort. Rock-solid stable. I enabled EXPO and away I went.
I love me some HUB! Which kit does he specifically use? I've got a G.Skill kit in my AM5 build so I'm curious.
Posted on Reply
#47
trparky
Gmr_ChickI love me some HUB! Which kit does he specifically use? I've got a G.Skill kit in my AM5 build so I'm curious.
G.Skill Trident Z5 NEO F5-6000J3038F16GX2-TZ5N.
I posted a link to the YouTube video at the specific timestamp where the memory kit's model number is in clear view. I can cite several videos (about two or three) where Steve mentions that specific memory kit.
Posted on Reply
#48
Gmr_Chick
trparkyG.Skill Trident Z5 NEO F5-6000J3038F16GX2-TZ5N.
I posted a link to the YouTube video at the specific timestamp where the memory kit's model number is in clear view. I can cite several videos (about two or three) where Steve mentions that specific memory kit.
Damn, couple numbers off on mine. Mine is the F5-6000J3636F16GX2-TZ5NR kit.
Posted on Reply
#49
trparky
Gmr_ChickDamn, couple numbers off on mine. Mine is the F5-6000J3636F16GX2-TZ5NR kit.
Your CAS latency is 36 whereas mine is 30. How much that means in terms of performance, I have no idea. Theoretically speaking, a lower CAS is better.
Posted on Reply
#50
csendesmark
trparkyYour CAS latency is 36 whereas mine is 30. How much that means in terms of performance, I have no idea. Theoretically speaking, a lower CAS is better.
Minimal, in practice you will not feel, but if you run only benchmarks, it can be measured :D
And yes, lower is better.
Posted on Reply
Add your own comment
Nov 24th, 2024 10:39 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts