Heisst, damit kann man einen oberen Grenzwert für das normale Core Performance Boost setzen ?oder ein negativen Offset auf max Boost beim Boost override (-1000mhz)
Follow along with the video below to see how to install our site as a web app on your home screen.
Anmerkung: this_feature_currently_requires_accessing_site_using_safari
Heisst, damit kann man einen oberen Grenzwert für das normale Core Performance Boost setzen ?oder ein negativen Offset auf max Boost beim Boost override (-1000mhz)
Welches Mod BIOS?@Reous
AGESA 1205 – Google Drive
drive.google.com
Hier sind alle Files enthalten, habse auch runtergeladen.
Owner hat es für das unify X schon drauf, gibt ein paar nette Änderungen.
Vddg für jedes ccd einstellbar z.B. oder ein negativen Offset auf max Boost beim Boost override (-1000mhz)
Ich flashe gerade für das x570 unify ein Mod BIOS.
Ok i got everything now@Veii
Thanks for the info.
Have you compared the temperatures of B0 to B2? I am especially interested in the max CCD temperatures. Run the b2 cooler?
prepared to depend on OC tools like Project Hydra
Ist hier ähnlich. Noch stabiler kann es nicht laufen.Mein B550 Gaming Edge Wifi läuft immer noch mit 1.2.0.3b.
Never change a running system.
Mod BIOS ist aus dem RAM OC discord, gemodded durch Boris (Eder ist dort auch vertreten und modded auch oft)
Ace - > https://cdn.discordapp.com/attachments/578340164187979796/916413943927427102/MEG_X570_ACE-1G5.zip
B550 Unify-X -> https://cdn.discordapp.com/attachments/578340164187979796/916409170360676402/B550UNIFY-XA45.zip
X570 Unify -> https://cdn.discordapp.com/attachments/578340164187979796/916410845972561950/X570UNIFYAB4.zip
Und ja, je nach Einstellung kann man noch das ein oder andere % Leistung herausholen, da die hidden Settings freigeschaltet werden. (Advanced -> AMD CBS -> L2 Prefetcher z.B.)
Instabil ist es bislang nicht (aber sicher ist halt noch nichts, da der Testzeitraum zu kurz ist)
@KedarWolf
Discord - Group Chat That’s All Fun & Games
Discord is great for playing games and chilling with friends, or even building a worldwide community. Customize your own space to talk, play, and hang out.discord.com
Interesting. But how come that the APUs don’t have them. The Specs are so similar I assume the just copy pasted the cores. If they fix it for them wouldn’t they have fixed it for b2 as well?Ok i got everything now
It's, too early to judge
And it's unfair to judge
I know now clearly what causes WHEA's and it's "unfixable" in the current state ~ and AMD likely will never fix it by the behavior their moderation team shows. An engineers mistake on the boost system , coming from Matisse days where they hardlocked dynamic SOC, MCLK, FCLK. They forgot to stop FCLK from beeing dynamic and overdriving itself for retail samples, resulting in potential#19 & #20 WHEA's @ & after 1800 FCLK, but not crashes (if one CCD is loaded too lightly by core affinity)
Also that B2 wasn't much different than B0
It remains speculation if substrate runs the same that XT will be on
But it is more powerefficient
Downside, it's far higher overvolted (in the sense that it takes what it wants correctly, but doesn't work well with current boosting system and allowed ranges)
Soo this results in a far hotter CPU, 220-250W with an Alphacool Pro @ 2500RPM & KPx pegs at 90c on y-cruncher AVX2 (stock PBoost)
Thermal density is tighter, but there is zero information if this is 7nm or smaller.
Just again, unfair to compare a Bronze vs Platin sample (both B2)
I wouldn't go out of my way to sell B0 anymore just to play silicon lottery again
Here you can either strongly lose or big times win, but both CCDs being Silver/Bronze and questionably "usable" in the current bios state. Too much efford
Thats my conclusion
PSP & Security Flaws could've been fixed on HW level
Potentially substrate is different , but it doesn't matter because:
A.) 1205 is not out till likely mid/end december
B.) AMD will 98% never fix their DPM issues and just have it fixed on the XT lineup (hopefully)
It goes tomorrow back to the client, i've had enough fun with it
With Hydra its fine, but current PBO is just not usable. It needs at absolute least -60 CO to balance it "on stock" , else it sits at barely 4.5-4.55Ghz by voltage throttle via FIT
Not to think about modifying CPPC by changing fMAX - then maybe the requirements are ~ -80 to -90 CO
Thermal throttle is at 90c, unchangeable so far
It runs hotter and feels denser packed
Nothing special in access latency, nothing special in memOC
AMD did hold their words, but potentially it can be good with 1205 ~ we'll see
EDIT:
Soo if you want to buy one
Be prepared to big times lose in silicon lottery and either lapp it , as it is hotter than a 5950X B0
Also be prepared to depend on OC tools like Project Hydra , as PBO is not much usable in the current state & most of the AMD OVERCLOCKING menu doesn't respond/isn't applied
Else i wish you best of silicon lottery luck. So far it has shown that either you win , or you lose big times
My personal advice is, to not sell away a "gold" rated B0 sample from an older batch & wait for the community to finally fix these DPM issues (potentially possible now with a clear target)
* well or just wait for XT
It's "sample research in progress" in this stateInteresting. But how come that the APUs don’t have them. The Specs are so similar I assume the just copy pasted the cores. If they fix it for them wouldn’t they have fixed it for b2 as well?
Can you give us some context how reliable that info is?
Thanks for dropping so much wisdom, do you mind if this finds it‘s way into the whea 19 discussion thread on overclock.net?It's "sample research in progress" in this state
There is still digging in progress to figure out what this toggle is actually called
Soo to see if it can be SMU disabled or it's deeper than userspace
They are remaings from the ABPDIS functionality - which where left alone
I feel the PBO X1-X10 ranges are also slightly broken (globally) but that's another topic
About "reliability" ~ there is no fix, to call it a news
The target/reason just got found, but not links to what feature does it
Just the problem part of it which is visible
Sadly also ZenPTMonitor Tool dev XPEHOPE, did not publish/release his tool in his github yet
Soo people need to dig and use pre-alpha releases to confirm their sample is not affected (well or just read whea log)
If the correct SMU value is found (the name of it)
Then we can forcefully annoy AMD , to finally fix their mess
So far , it's just an error found, without a resolve. No pressure can be made on a found error.
But yes, it's reliable about what actually does trigger it, just not the connection to it what triggers this overdrive
* as some units don't have this and where last minute patched.
EDIT ~ Followup,
Soo issue for FCLK_EFF high readout (boost)
Is CSTATE_BOOST triggering on Idle. Overdriving beyond set null-point
This behavior depends on the OS Scheduler (the amount of behavior)
Win11 in specific.
Performance mode under Ryzen Powerplan does boost on mousemovement specific. Balanced does only boost on idle
Battery Saver boosts correctly like Win 10, and doesn't also increase DPM LCLK beyond 593mhz
DPM LCLK limits "have to be" 770mhz peak" , this goes for RX Cards too
Anhang anzeigen 699644
But their 619 ~ 310-619 LCLK strap, is "problematic" - fixing it to 619-619, or rather 593-593 , removes the LCLK_EFF back and forth shuffling
AMD CBS does only allow to set DPM 1 = 300mhz , DPM 2, = 600mhz , but not DPM 3 = 770mhz, yet PCIe 4.0 cards try to reach this 700+ mhz
Anhang anzeigen 699643
This can be fixed by fully disabling DPM LCLK , but you lose PPT throttle (i don't think anything bad so far).
On "fixed samples" you also lose THM tracking together with procHot - on our retail you just lose PPT but prochot and THM remains active
* or via Mailbox on boot, but load-balance is still active
~ disabling DPM LCLK removes part of the big problem from the equation fully
Active issue remains still that CSTATE_BOOST on idle overdrives, soo takes FCLK_EFF with it
Meaning up to how you set your sheduler & powerplan to behave ~ you will get these WHEA's by different "types" / "methods"
On stock, you get them always on low idle, and nothing on load
On Perf-Mode, you get nothing on idle, but it can overdrive on load ~ and so on
More to follow, maybe in another thread
But as Bios update information
RX users (with only one NVMe) can fixate the DPM link of their cards to 593-593 or 613-613 with MorePowerTool
~ and Win11 users put powermode to balanced powerplan with energy efficient target. That's "normal" and without random overdrive spikes
(this all includes the overboost bug that reaches 5500mhz and higher)
Soo expect DPM lv 3 support (maybe) ,
~ and if not - might want to lower RX cards DPM level down to 593-593, or 319-593 , else 594-770 if DPM 3 get's allowed via SMU/Mailbox finally/again
If security vulnerabilities would be fixed ~ in this case "SMU RACE" then it also introduced ROM ARMORMajor vulnerabilities updates, customers are strongly encouraged to update to this release at the earliest.
Credits to "Assaf Carlsbad and Itai Liba from SentinelOne"
• Introduce capsule BIOS support starting this version.
As demand to allowing targets via userspace to excecute and rewrite ROMSome people think having unrestricted writes to SPI flash is OK:
AGESA FW stack patched bioses for 3rd gen
www.overclock.net
I'm already crosslinking forums on important updates. Don't worrydo you mind if this finds it‘s way into the whea 19 discussion thread on overclock.net?
Please add this warning as * under the F51c fileI can not out of clear mind say, downgrade back would be possible after the flash
Haben die neueren bzw zukünftigen UEFIs einen Vorteil, gar überhaupt einen Sinn auf Ryzen 1000 ?
Bis auf allgemeine Fehlerkorrekturen wird sich da nix mehr tun. Warum sollte AMD an ZEN1 rumbasteln, wenn sich vorher schon keine großen Probleme mehr gezeigt haben.Ich habe bereits mit dem Ryzen SMU Checker die einzelnen SMUs der Nachfolge Versionen des BIOS einmal durchgecheckt und habe festgestellt, dass sich bei Summit Ridge nichts mehr tut.
Okay danke schonmal dafür.Bis auf allgemeine Fehlerkorrekturen wird sich da nix mehr tun. Warum sollte AMD an ZEN1 rumbasteln, wenn sich vorher schon keine großen Probleme mehr gezeigt haben.
Ein Bios Update ist keine Pflicht, nur eine Option zur Verbesserung bei großen Problemen.
Nur wenn du mal neuere CPUs einsetzen solltest, würde ich vorab nen Bios Update machen, denn sonst gerät man schnell in die Falle, wenn dann keine ältere CPU mehr vorhanden ist.
Wenn sich nicht irgendetwas wie Spectre, Meltdown o.ä. und eine gewisse Anfälligkeit bei Zen1 zeigt, wird AMD wohl nix mehr dort ändern.
Das wäre für mich der einzige Grund, den AMD noch für Änderungen hätte.
Wenn es für Zen1 neue SMU Versionen gibt sollten die Optimierungen auch da greifen.Bis auf allgemeine Fehlerkorrekturen wird sich da nix mehr tun. Warum sollte AMD an ZEN1 rumbasteln, wenn sich vorher schon keine großen Probleme mehr gezeigt haben.
Ein Bios Update ist keine Pflicht, nur eine Option zur Verbesserung bei großen Problemen.
Nur wenn du mal neuere CPUs einsetzen solltest, würde ich vorab nen Bios Update machen, denn sonst gerät man schnell in die Falle, wenn dann keine ältere CPU mehr vorhanden ist.
Wenn sich nicht irgendetwas wie Spectre, Meltdown o.ä. und eine gewisse Anfälligkeit bei Zen1 zeigt, wird AMD wohl nix mehr dort ändern.
Das wäre für mich der einzige Grund, den AMD noch für Änderungen hätte.
Wollte gerade mal nach neuen BIOS Updates schauen und als ich die MSI Webseite aufgerufen habe, die englische und nicht die deutsche, kam dieses hier:
Anhang anzeigen 703127
Da ist wohl was schief gelaufen oder die sind am Umstellen.