AMD Zen 2 (Ryzen 3000) - *** NO COMPETITOR HINTING ***

[MegaPinches]

I'd argue that the board is faulty, as every issue you've described are things that should not happen. Even if this is down to bad BIOS, it's still faulty however you slice it and as a consumer you shouldn't have to wait for a fix to come out via a BIOS release if it renders the PC unusable.

I would speak to OCUK and see if you can return it and get a refund to trade up to a mid range x570 or x470 (I'd go x470 if prices had come down... but they don't appear to have budged).

I just don't wanna seem the type returning anything that has a speck of dust in the wrong place you know?, i'm not like that, ive just returned two LG monitors as they both had the same big fault i'm gonna look the type and like I said earler i'm confident its not user error, ive been building pc's for 18 years now and it couldnt be easier now but this is a non stop hassle. I might speak to them then. Not looked too greatly into x570 can you point me in direction of the goto lower priced boards with decent VRM's. I read recently 3 models have already been disountined for being poor and replaced but i can't remember what they were.

 

[KungFuSpaghetti]

I just don't wanna seem the type returning anything that has a speck of dust in the wrong place you know?, i'm not like that, ive just returned two LG monitors as they both had the same big fault i'm gonna look the type and like I said earler i'm confident its not user error, ive been building pc's for 18 years now and it couldnt be easier now but this is a non stop hassle. I might speak to them then, hopefully I can order new one first so no down time. Not looked too greatly into x570 can you point me in direction of the go to lower priced boards with decent VRM's. I read recently 3 models have already been disountined for being poor and replaced but i can't remember what they were.

Not sure on which x570 boards to go for to be honest, probably best asking in the mobo section as I've done little to no research into them. I do hear generally that the Gigabyte and Asus ones are good though and the boards that were considered hot garbage on 400 series for Asus, like the TUF and PRIME are now actually pretty decent. Only thing I will say is that I'd check where the chipset sits on the board, as some look to be under where a GPU will sit.

 

[MegaPinches]

Not sure on which x570 boards to go for to be honest, probably best asking in the mobo section as I've done little to no research into them. I do hear generally that the Gigabyte and Asus ones are good though and the boards that were considered hot garbage on 400 series for Asus, like the TUF and PRIME are now actually pretty decent. Only thing I will say is that I'd check where the chipset sits on the board, as some look to be under where a GPU will sit.

Ok cheers for advice gonna have a gander later, gotta get back to work.

 

[peterwalkley]

To confirm others' experience: I've tried the 7C02v34 release (25th October) of the MSI Tomohawk max BIOS and found it very problematic. Constant BSOD's and this was without XMP enabled or any form of overclocking - whole thing at stock and not running any form of stress tests or significant load (just catching up on emails after being away a week).

I've rolled back to the earlier 7C02v32 release (2nd September) just to get stability back.

 

[alec]

Well according to this video anyway.... everything may not be so peachy with the new 10.0.4 Agesa.

Offtopic, but I hate getting information via video. I'm sure the issue can be summarised in a tweet sized message, but no, have to go and see a dude talk about it for unnecessarily long time

 

[AStaley]

For those of us at work who've not been able to check the video yet, what's the issue with the 1.0.0.4 update?

 

[Woodsta888]

Offtopic, but I hate getting information via video. I'm sure the issue can be summarised in a tweet sized message, but no, have to go and see a dude talk about it for unnecessarily long time

There's no talking on it. You gotta watch and look at all the numbers as the video goes through a load of benchmarks.

For those of us at work who've not been able to check the video yet, what's the issue with the 1.0.0.4 update?

He/she keeps stating at the start of each bench that his settings are the same. But he see's higher frequencies but lower FPS. They test a 3600X 3700X and a 2700X

 

[humbug]

For those of us at work who've not been able to check the video yet, what's the issue with the 1.0.0.4 update?

He's getting about 5% less performance 1004 vs 1003 ABBA in games, including AC Origins, other than CS:GO where its about 5% higher.
This with an MSI board.

His previous test, he mentions this in the description, was with an ASRock board and he got 10% higher performance in AC Origins.
Looks like MSI have some more work to do.

AC Origins ASRock board. +10% vs 1003 ABBA.
https://youtu.be/0MWsrMHp5j4?t=175

AC Origins MSI board. -5% vs 1003 ABBA.
https://youtu.be/2qy_IeUZqxU?t=328

The ASRock board might even be a little more than +10 vs 1003 ABBA, looking at it again... so it looks like the MSI board running on 1004 is 15 to 20% slower than the ASRock board on the same AGESA

 

[Buffalo2102]

The frame loss is minimal and probably unnoticable. What he fails to mention (notice?) is that with the 1.0.0.4 Agesa the CPU power is typically a third of the Agesa 1.0.0.3 and often even less than that. Seems a pretty good trade-off to me.

 

[Woodsta888]

The frame loss is minimal and probably unnoticable. What he fails to mention (notice?) is that with the 1.0.0.4 Agesa the CPU power is typically a third of the Agesa 1.0.0.3 and often even less than that. Seems a pretty good trade-off to me.

He's getting about 5% less performance 1004 vs 1003 ABBA in games, including AC Origins, other than CS:GO where its about 5% higher.
This with an MSI board.

His previous test, he mentions this in the description, was with an ASRock board and he got 10% higher performance in AC Origins.
Looks like MSI have some more work to do.

AC Origins ASRock board. +10% vs 1003 ABBA.
https://youtu.be/0MWsrMHp5j4?t=175

AC Origins MSI board. -5% vs 1003 ABBA.
https://youtu.be/2qy_IeUZqxU?t=328

The ASRock board might even be a little more than +10 vs 1003 ABBA, looking at it again... so it looks like the MSI board running on 1004 is 15 to 20% slower than the ASRock board on the same AGESA

That's a very good point. I hadn't considered which MOBO manufacturers BIOS he was testing or that they'd be different results between them.

 

[humbug]

The frame loss is minimal and probably unnoticable. What he fails to mention (notice?) is that with the 1.0.0.4 Agesa the CPU power is typically a third of the Agesa 1.0.0.3 and often even less than that. Seems a pretty good trade-off to me.

Its really not....

First image looks like its at 720P: (MSI Board)
ABBA 165 FPS (+6%)
1004 153 FPS

Second looks like its 1080P (ASRock board)
ABBA 130 FPS
1004 152 FPS (+17%)

That's a whopping 23% difference between boards.

 

[Buffalo2102]

Yes but are you going to actually notice the difference between 130 and 152 FPS at 1080p. I think very few will.

 

[humbug]

Yes but are you going to actually notice the difference between 130 and 152 FPS at 1080p. I think very few will.

Agreed. But, IF 1004 increases gaming performance by 15% in AC where MSI actually get a small regression in performance because they messed it up i think it really matters, AC is one of those 5 or so out of 36 of HUB games where Ryzen 3000 is 10%+ slower than Coffeelake, with these numbers a 3800X would be faster in AC Origins than a 5Ghz 9900K.

 

[Panos]

@humbug have tried to run the benchmarks stock without changing anything like PBO?
Just with XMP only.

 

[Buffalo2102]

No argument here.

But maybe I misunderstood the point of the video. I thought he was trying to point out an issue with 1004, not an issue with MSI's implementation of it. Or am I still not understanding?

 

[humbug]

@humbug have tried to run the benchmarks stock without changing anything like PBO?
Just with XMP only.

PBO makes 0 difference in games, tho i think that because i'm still on 1003 ABB, Mhz seem to be hard locked to 4.125Ghz.

It does make about a 100Mhz difference in hard MT loads.

I have benchmarked a couple of games XMP 3000 vs DRam Clac tuned 3333 and i get about +10%

 

[Twinz]

I'm on and MSI x570 carbon. My CB R20 MT score went up by 10 points. Time-spy went down a little from 112xx to 111xx and my shadow of the tombraider bench went up by a few FPS on all the CPU numbers.

The thing I noticed is that the run-to-run variance is tighter than before. ABBA was capable of putting up good numbers but seemed to vary more from one run to the next. I wonder if I average 10 runs of TS if the latest BIOS wouldn't come out ahead in that bench too.

I haven't really dug into yet though. I want to spend some time in Project Cars 2 for a while (my rig's primary purpose) to see what I think of the new BIOS. (If I even notice a difference at all)

 

[humbug]

No argument here.

But maybe I misunderstood the point of the video. I thought he was trying to point out an issue with 1004, not an issue with MSI's implementation of it. Or am I still not understanding?

Yes, i think he's missing the nuance of the fact that they are different boards, given the performance boost on the ASRock board vs the performance regression on the MSI board on the same 1004 AGESA it would suggest its a problem with the MSI board, not the 1004 AGESA.

I put that to him, he just replied to me saying that the ASRock board was actually 1003 ABB (not 1003 ABBA) vs 1004, 1003 ABBA was only to fix boost issues, you don't get 15% higher performance from a 50Mhz higher clock, i put that back to him.

 

[Buffalo2102]

Yes, i think he's missing the nuance of the fact that they are different boards, given the performance boost on the ASRock board vs the performance regression on the MSI board on the same 1004 AGESA it would suggest its a problem with the MSI board, not the 1004 AGESA.

I put that to him, he just replied to me saying that the ASRock board was actually 1003 ABB (not 1003 ABBA) vs 1004, 1003 ABBA was only to fix boost issues, you don't get 15% higher performance from a 50Mhz higher clock, i put that back to him.

OK cheers humbug.

 

[Panos]

PBO makes 0 difference in games, tho i think that because i'm still on 1003 ABB, Mhz seem to be hard locked to 4.125Ghz.

It does make about a 100Mhz difference in hard MT loads.

I have benchmarked a couple of games XMP 3000 vs DRam Clac tuned 3333 and i get about +10%

Sorry was a benchmark

However many pointed out at him that his numbers are getting contradicted with a previous video.