• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

AMD TLB Notes and AOD Settings

FrgMstr

FrgMstr

Just Plain Mean
Staff member
2FA
Joined
May 18, 1997
Messages
58,239
As we mentioned last week, AMD has fixed its TLB issues in upcoming silicon. We are still getting questioned on this, so here is an official statement on that direct from Damon Muzny at AMD PR.

News Post Here
 
Some of the "readers" of this forum are quite the users, under normal circumstances, many will not encounter this bug. How many of us here are what you would consider normal users though?

Also it seems out of the box, it will now be quite the performance cripple against it's price rival from intel. Sure you can disable the fix, but with AMD not even hosting it... seems kinda counter productive. I mean, if you aren't worried about shutting off the fix and having a stable system regardless, why not host the file and make it easier for users to get to it?
 
As I understand it you should be able to get it from your motherboard makers websites. I mean how many poeple go to AMD for the drivers and utilities anyway, I have always looked to the motherboards maker website first for those items. this doesn't seem like a problem to me, nor any kind of deviation from normal protocols for software distribution from AMD. I could be really wrong here, as I have never gone to AMD looking for any software.
 
So what would you really have to do to ever experience this TLB erratum? Be running virtualization? Because 100% CPU Utilization isn't that hard... we ALL do it to test stability of our systems with Orthos, etc. Stressing the CPU is a common part of the enthusiasts daily usage of the computer.

But I also remember you saying Kyle that you've been using a Phenom without the fix and haven't experienced the problem, along with others who now own Phenom's. If I ended up getting one, I probably wouldn't bother turning it on unless I somehow experienced it.
 
HardDrugsUser said:
Don't underestimate the users, i will be able to create that load.
Click to expand...

i dont think it was a challenge or a belittling remark. i think most users would probably not encounter an issue with this errata. it reminds me of the original Pentium snafu intel had with certain fp operations. people heard of the errata and demanded it be fixed/replaced.

knowing how to trigger the scenario where the TLB errata would compromise a system is wholly different from the average user or even a enthusiast encountering the issue via normal usage. although i do think many people underestimate [H]'rs and truth be told the cumulative expertise and knowledge among us "enthusiasts" is impressive.

i am glad AMD has addressed the issue and has taken the time to clarify the situation.

now how about Bulldozer? :rolleyes:
 
Thanks for the heads up Kyle. I wonder why some people like to make this issue bigger than it really is, some even went so far by spreading false rumour about B3 stepping still has the bug.
 
I don't give a shit about the Phenom issue, but AMD is really screwing the pooch with the Opteron. We've held out for Opteron QCs for as long as possible, so it looks like Intel will be running in our VMWare boxes.
 
Would be nice if the B3 stepping also brings in some other improvements, like higher speeds and performance.
 
Thanks for the front page news Kyle. There has been a lot of misinformation about the effect of the TLB erratum. The Lost Circuits article did a good job explaining the technical details. I will reiterate that the erratum IS fixed in B3.
 
Thanks Kyle, wootage! :cool:

alg7_munif said:
Thanks for the heads up Kyle. .
Click to expand...
The above quote is what should have been said.

alg7_munif said:
I wonder why some people like to make this issue bigger than it really is, some even went so far by spreading false rumour about B3 stepping still has the bug.
Click to expand...

You're making the issue bigger than it really is, you're doing so right now! Stop putting AMD in such a bad spot, learn to accept the facts and be done with it.
 
HardDrugsUser said:
Don't underestimate the users, i will be able to create that load.
Click to expand...

Uh, I seriously doubt it. In fact, I will pay you $1000 if you can recreate it on a non-virtualized desktop system and prove it to me.


And we are going to change your forum nick. Your current one is not acceptable. PM me with what you would like it changed to please.
 
Kyle can you confirm whether there would be any other improvements other than the tlb fix? higher clockspeeds, etc etc?

and this is indeed interesting, check

a glimpse of hope

hopefully as games and programs become more multithreaded, we will see more gains

and to the intel camp: i am not (i repeat:NOT) saying agena>core. Thank you
 
RangerSVT said:
Thanks Kyle, wootage! :cool:


The above quote is what should have been said.



You're making the issue bigger than it really is, you're doing so right now! Stop putting AMD in such a bad spot, learn to accept the facts and be done with it.
Click to expand...

honestly, stop trolling in all of the amd threads. alg7_munif said nothing wrong in his post:confused:
 
I've been running the red selection myself without any issues. I am currently just doing a flat 2.4GHz though. Left on its own my system defaults to green with the yellow circle about it, or the middle option.

It's funny how many folks that don't own and haven't tested a Phenom have much to say about the "TLB bug".
 
does the new core with the tlb fix still suffer the performance penalties? or is it as good as the old core with the tlb fix disabled?
 
ziddey said:
does the new core with the tlb fix still suffer the performance penalties? or is it as good as the old core with the tlb fix disabled?
Click to expand...

Nobody knows this yet, but the idea is that it would be as fast as without the fix. I'm sure there will be a number of tests comparing the new processors with the old processors at the same clock frequency.

More on topic: In all fairness, I think Phenom and AMD overall has begun to become irrelevant to enthusiasts. Maybe it's just me, but I used to be a hardcore AMD fanboy and I'm going to pull the trigger on a Q9450 in the next month because AMD simply misrepresented the performance of Phenom numerous times before its release. Hell, I even stuck with them and bought an AM2 rig to tide myself over until the release of Phenom even though Core 2 Duo was a much better enthusiast platform. God knows if AMD can become competitive again in the next few years, but I think too many people are clinging to the idea that Phenom can still do it.
 
I think AMD is kind of being wishy washy on this. Its not a big deal that won't affect you, but we're putting something in place, that you can disable, IF you find it somewhere on the web yourself.

Additionally, please note how the three “Turbo” button settings in AMD OverDrive correlate regarding disabling the TLB erratum fix.

Green (top image) = Safe performance level (no TLB caching / default for B2 CPU)

Yellow (middle image) = Elevated performance level (TLB caching enabled… TLB erratum fix turned off)

Red (bottom image) = Highest performance level (TLB enabled + CPU pwr management tweak enabled)

Note: the TLB erratum fix is disabled by moving above the Green / Safe performance level… done so at the user’s discretion, of course.
Click to expand...

They're covering their asses by warning you about it and not officially hosting it or distributing the AOD, which to me means its more likely to happen than they're admitting.

But I guess the point is kind of moot anyway, as why would I buy a slow broken processor (even if its only 0.01% broken), when I can buy a perfectly good working processor thats faster at stock (clock for clock) and comes in much faster speed grades, and overclock much easier and higher.
 
You must log in or register to reply here.
Back
Top