Ryzen has another instruction implementation problem, VME is broken

[Phate.exe]

15 hours ago, MrDynamicMan said:

Any company/org still running such old OS would not be upgrading to brand new pcs without upgrading OS infrastructure. The only place I can see this being an issue is if the tech dep got new PCs and emulates older OS's for testing. Which is a reallly, realllllllly, reałllllllllllllly (lol) neich case anyways. Sure, AMD might lose the business from those <100 companies, but its not worth the RnD for them to fix it.

In Pharma and Biotech facilities, there is a ton of ancient equipment that requires some archaic software to interface with.

 

The haswell laptop I use has a bunch of VM's on it just so I can use said old software.  The WinXP VM is the one I use most often.

[zMeul]

2 minutes ago, jasonc_01 said:

Thats exactly what he SAID! Quote "They arent available, that was the point. Jesus" and you replied with Quote "geez .. they are not! good god!". 

 

Dude this isn't hard to comprehend, you corrected him with his own answer! 

you should follow the thread and not just the last few posts, start at the top of this page

 

good god! you people have nothing better to do? you're not even following the discussion

[leadeater]

2 hours ago, BaRoMeTRic said:

were not talking server platform, we're talking consumer grape chips here. so if AMD has a few bugs for a 5 gen old os, on consumer platforms it's not that big of a deal. If this happened on the new server lineup then it would be a HUGE deal

 

2 hours ago, zMeul said:

do you have any idea how VMs work? apparently not

 

consumer?!

please tell me the difference between the consumer Ryzen and the enterprise Ryzen version

 

1 hour ago, leadeater said:

Naples and Snowy Owl are the enterprise products not Ryzen.

 

Naples:

• More CCX modules
• 64 PCIe lanes
• 8 Channel memory
• Socket SP3 (4094 pins)
• LGA not PGA

Still these all use the Zen architecture so it's likely this bug is in the Naples/Snowy Owl engineering samples and will be fixed before launch. Good to know that the advice I was giving to not use Ryzen for virtual hosts was correct.

 

 

1 hour ago, leadeater said:

True but there are no Ryzen servers other than custom built ones out of parts which no sensible business would do and as I stated I've already been advising not to use Ryzen for VM hosting.

 

1 hour ago, zMeul said:

please point to me where the server grade Zen based CPUs are available

 

1 hour ago, Clanscorpia said:

They arent available, that was the point. Jesus

 

Bolded for ease of reference and includes relevant and important posts. Seems clear to me the discussion chain started with talking about the server platform (not the OP/Topic), whether it existed or not wasn't the point. During the discussion it was clearly noted that no server platforms have been released yet.

 

@jasonc_01

Please tone it down a little, it's not acceptable to suggest or imply someone needs to take medication. 

[XenosTech]

8 minutes ago, leadeater said:

-snip-

Well that was a low blow lel

[LAwLz]

3 hours ago, Energycore said:

I think you're using inaccurate data here. You're saying that only 10% of topics posted have been negative towards AMD, but that number doesn't help at all because every topic that isn't about AMD can't be negative toward the company anyway (not unless it's off-topic bashing which is ill advised).

 

I think the number that matters is that 3 out of 8 topics about AMD were negative, which constitutes 37.5% of all topics.

I think that's a really biased way of looking at those statistics. So if I make 1000 topics, and only one is about AMD and that's negative, then 100% of my posts has been negative posts towards AMD and people can ridicule me for it? Come on... Here is the sampling bias again.

You're not looking at all his posts. You're discarding 75% of his posts just to artificially increase the number of "bad post percentage" he has about AMD. You take all the topics he has created in the last months, not a subsample of them.

 

You're simply wrong. Not 37.5% of all his topics are negative news about AMD. 10% of all his topics in the last 2 months have been bad news about AMD.

It is impossible that 37.5% of his topics would have been negative things about AMD, because only 25% of his topics has been about AMD at all, most of which has not been negative news.

 

 

3 hours ago, Energycore said:

I won't comment on whether that reveals a bias or not. But I also believe you're not being objective in that analysis even though you claim to, because labeling a topic positive / negative is a process that requires a value judgment which can never be objective. It would be objective to say that "X topics out of Y about AMD included one of the following negative words directed toward AMD: incompetent, mistake, problem, broken (...)".

 

So if I may, how did you choose which topics were negative, neutral and positive?

I used common sense and reasoning.

Something like "Windows 10 update improves performance of Ryzen processors". Who in their right mind would say that's a negative thing about AMD? It is clearly positive, right?

Meanwhile, I don't think anyone would argue that this topic, that there is a bug with Ryzen processors is anything but negative. Correct?

Of course there is no mathematical formula to determine if something is good or bad news for AMD, but there is a subjective opinion which 99.9% of people will agree with. That's what I used on that particular point.

[zMeul]

40 minutes ago, LAwLz said:

-

why does it even matter?!

 

I could be writing 1000% of the time only negative topics and posts about AMD

what would matter would be whether if the content is or not valid, researched or back-up with sources

are my posts backed-up? yes they are - then why do we even need to discuss this?

doesn't sit well with RedTeam+? they could go eat a truckload of rocks for all I care

 

oh, and PS: I do have something against AMD 

 

[leadeater]

1 hour ago, zMeul said:

why does it even matter?!

 

I could be writing 1000% of the time only negative topics and posts about AMD

what would matter would be whether if the content is or not valid, researched or back-up with sources

are my posts backed-up? yes they are - then why do we even need to discuss this?

doesn't sit well with RedTeam+? they could go eat a truckload of rocks for all I care

 

oh, and PS: I do have something against AMD 

 

Actually most of the topics you create aren't really, or at all, that negative and you're one of the only people I know of that goes back and updates them when new information comes out, even if the topic is rather old.

 

You may have a negative slant on things during discussion in the topics but it's not anyone's place to say you can't do that, your free to have your own opinions and preferences. Since this is a discussion forum everyone has the right of rebuttal, not ridicule.

[TOMPPIX]

1 hour ago, zMeul said:

oh, and PS: I do have something against AMD 

could you please elaborate on that, so we can understand you better.

[zMeul]

Just now, TOMPPIX said:

could you please elaborate on that, so we can understand you better.

I did it some time ago and I got bored explaining it again and again

simply put it, AMD fucked my wallet and I will fuck with theirs by exposing every shit they do

[TOMPPIX]

5 minutes ago, zMeul said:

I did it some time ago and I got bored explaining it again and again

simply put it, AMD fucked my wallet and I will fuck with theirs by exposing every shit they do

so your intentions is to make amd look bad, because of something that happened some years ago.

[zMeul]

Just now, TOMPPIX said:

so your intentions are to make AMD look bad, because of something that happened some years ago.

I'm not making AMD look bad, they do that by themselves 

[TOMPPIX]

Just now, zMeul said:

I'm not making AMD look bad, they do that by themselves 

true

[zMeul]

some just told me AMD might've fixed the VME bug in the latest AGESA, but I can't find any source to confirm it

anyone has any idea on it?

 

I only found this: https://www.heise.de/newsticker/meldung/AMD-Ryzen-Kommende-BIOS-Updates-patchen-auch-VME-Bug-3732056.html

but it's not a concrete confirmation

[leadeater]

6 minutes ago, zMeul said:

some just told me AMD might've fixed the VME bug in the latest AGESA, but I can't find any source to confirm it

anyone has any idea on it?

 

I only found this: https://www.heise.de/newsticker/meldung/AMD-Ryzen-Kommende-BIOS-Updates-patchen-auch-VME-Bug-3732056.html

but it's not a concrete confirmation

It was confirmed on AMD's website or blog, I forget which. There is a news topic about it.

 

Edit:

On 5/8/2017 at 11:28 PM, Morgan MLGman said:

For those looking for IOMMU fixes we are hopefully going to have an option to force boot off a specific PCIe slot. It's not the grouping fix, but a work around for now.

So a workaround rather than a fix.

[zMeul]

2 hours ago, leadeater said:

workaround rather than a fix.

none of those have anything to do with the VME bug

 

the IOMMU outstanding issue is a different problem

[leadeater]

Just now, zMeul said:

none of those have anything to do with the VME bug

 

the IOMMU outstanding issue is a different problem

Hmm I thought the VME bug and IOMMU were related, not paying much attention to them yet since no products I would want to buy are out yet.

[grimreeper132]

On 2017-5-13 at 6:29 PM, Darth Revan said:

So can we agree now that AMD launched a beta product?

Or do you guys still need more proof?

 

P.S. I know everyone is going to start the hate because AMD delivered a good performer. Yeah yeah yeah, that's not the issue. So start with that.

I agree with you to a point, it's past beta, but not by much, and it still has bugs.

That being said what they did launch even with it's bugs is far superior to intels offerings, at that price range, in many applications maybe not gaming but in many others it is.

[Notional]

This issue was solved in the newest AGESA version: https://www.heise.de/newsticker/meldung/AMD-Ryzen-Kommende-BIOS-Updates-patchen-auch-VME-Bug-3732056.html