AMD Phenom is BROKEN - Costs a 14% performance hit to fix

December 6 Carl Nelson

Subscribe to our FULL RSS feed!

We recently published our AMD vs. Intel Dual Core comparison, which led some to wonder why we didn’t compare their quad-core chips instead. As I stated in the review, one reason was that we simply wanted to write a CPU review discussing products that more people can afford; even the high-end dual-core processors we looked at were all under $200. The cheapest quad-core is in the $300 range, both for Intel and AMD. The other problem was that we simply didn’t have a Phenom to test. We have to buy our AMD processors for review, and I hadn’t gotten around to that. I originally wanted to wait until they got their 2.4 GHz part fixed, so I could test the full Phenom lineup.

How wrong I was.

As it turns out, the Phenom launch has been plagued with nothing but blunders, and it may go down in history as one of the worst PC hardware launches in history.

First of all, it wasn’t just the 2.4 GHz part that is affected by the erratum - it’s the entire lineup of Phenom processors, along with all Barcelona Opterons. Basically, the problem is with the CPU’s translation lookaside buffer (TLB) and L3 cache, and can cause system crashes and data corruption. The specifics are as follows:

Erratum 298 will be described as follows: “The processor operation to change the accessed or dirty bits of a page translation table entry in the L2 from 0b to 1b may not be atomic. A small window of time exists where other cached operations may cause the stale page translation table entry to be installed in the L3 before the modified copy is returned to the L2. In addition, if a probe for this cache line occurs during this window of time, the processor may not set the accessed or dirty bit and may corrupt data for an unrelated cached operation. The system may experience a machine check event reporting an L3 protocol error has occurred. In this case, the MC4 status register (MSR 0000_0410) will be equal to B2000000_000B0C0F or BA000000_000B0C0F. The MC4 address register (MSR 0000_0412) will be equal to 26h.”

The reason they were able to proceed with the launch is that this particular erratum is fixable via a microcode update (done using a motherboard BIOS update). Basically, the part of the L3 logic that causes the problem is disabled. However, running the fix instills a performance penalty, because of how the L3 cache’s throughput and latency is affected. There was no way to know exactly how performance would be affected, except for the “around 10%” estimate that has been floating around. Fortunately, our friends at Tech Report have obtained a retail Phenom 9500, and tested its performance before and after applying the patch on an MSI motherboard.

The results were even worse than expected. When looking directly at cache memory performance, bandwidth dropped by as much as 38.7%, and latency slowed down by over 50%!

Of course, what really matters is how the actual programs we use every day are affected. Tech Report ran a full suite of Worldbench tests, to find that the performance drop ranged from around 5% to as high as 50% in programs like Firefox. The average drop in performance in their tests ended up being 13.9%.

So if you bought a Phenom, you have a broken CPU. If you want to fix it, it will cost you. Not in cash, but in performance. It’s basically like paying for a 2.3 GHz CPU but really getting a 2.0 GHz.

It is expected that future revisions of the Phenom chip will fix this error in the silicon itself, so there should be no performance hit at all. If you were considering buying a Phenom processor, you might want to hold off until the next revision (B3) comes out, and is confirmed to work at full capacity.

But that’s not all.

What really got me irate was another finding Tech Report made when doing these tests. When they plugged in their retail boxed Phenom 9500 CPU, they noticed that the northbridge was clocked at 1.8 GHz, rather than 2.0 GHz on the engineering sample 9500 that was provided to them by AMD. This would be fine if AMD had mentioned that the northbridge is not running at retail speed, but according to Tech Report, they were told that 2.0 GHz was the correct speed for the Phenom 9500. Obviously, that is not the case, and once again we are seeing AMD try to ‘trick’ hardware sites into reviewing overclocked products as normal retail products.

The overclocked northbridge helped increase the Phenom’s benchmark scores across the board too, which artificially made the CPU look more competitive to Intel’s competing products.

It’s interesting to see AMD come full circle from when they blacklisted us for giving them trouble over launching ‘paper’ products to review sites that never made it to retail, to kicking Intel’s ass during the PressHot days, to launching a lineup of broken CPUs that not only perform 14% lower when fixed, but are slower than initial reviews indicated to begin with.

  • N4G
  • Digg
  • Facebook
  • Slashdot
  • StumbleUpon
  • TwitThis
  • Technorati

Filed under: PC Hardware

Tags: , ,

Related Posts:


Abit AX78 Review
A look at CPU prices and performance - in relation
PS3 folding performance isn’t THAT great
Tweak Vista Tip #2 - Enhance SATA Drive Speed
Corsair Flash Survivor memory can take a beating

14 Comments »

Comment by Trevor Flynn
2007-12-06 06:21:05

I’d be pretty pissed if I was an early adopter.

The overclocking review samples is sad too, but unfortunately not a giant surprise. Remember when Nvidia used to write specific driver code that lowered quality during 3DMark runs in order to attain higher scores. This at least is more easily detected, if not stupider on AMDs part. I can’t imagine they’d think no one would notice or call them on it (though I guess they though maybe no one check them NB bus speeds?).

More than likely I’d think it would be a yield issue and they had to downclock to get enough of them out the door for launch.

 
Comment by Goty
2007-12-06 13:17:23

So, no mention of the fact that the TLB erratum really only affect people using virtualization software I see.

 
Comment by Carl Nelson
2007-12-06 17:19:39

No mention, because it’s not a fact

 
Comment by Scott
2007-12-07 07:55:49

It’s really sad how far they’ve fallen. Very good article.

 
Comment by mike
2007-12-07 09:25:24

if you are gonna build a system, nows the time before intel hikes prices

 
Comment by Robert Subscribed to comments via email
2007-12-09 17:56:24

I’m running a Phenom 9600 on a MSI K9A2 and have seen no stability issues yet when running games like FEAR and Halo 2. I haven’t tried tried all my software out yet but I think this whole errata thing is being exaggerated a bit. I even have the thing watercooled down to as low as 17 - 18 degrees Celsius with a Nautilus 500 from Corsair. The chip works fine right out of the box at 2.3 ghz in Windows Vista 64 bit Ultimate. Unfortunately overclocking is not working at all despite the low temps. Alter the FSB and get nothing but crashes. Right now I don’t have a definite reason for that. They aren’t totally broken.

 
Comment by jason
2007-12-13 20:24:17

i think that amd should tell the truth about it i want some god blasted truth about this the processor seem to work fine for other people did you know that your actualy supposed to run amd phenom on the new amd 790 chipset toget full power i herd that amd is releasing revision b3 in 2end qourter 2008 around febuary or march

 
Comment by Dave Subscribed to comments via email
2007-12-19 08:14:18

I just bought a 9500, got a nice aftermarket zalman cooler so i could over clock it….thing crashes with a small FSB increase by a measly 10mhz. This processor absolutely blows, I bought it thinking the 9600 was the one with the issue. I’ve always been a AMD loyalast…but now I shoulda listened and went with intel, what a waste of $$$ (I’ve also heard that microsfot has a patch which has a 0% penalty…is this true? sounds like b.s. to me) Anyone else have the same problem overclocking a phenom…you can’t even if the CPU temp is low?

 
Comment by Robert Subscribed to comments via email
2007-12-20 18:24:10

Until we are sure it is the processor I think we should reserve judgement. I know that I am currently having a memory issue where with the current BIOS I have my system memory is running at a voltage below the 2.2 volts stated by Kingston to operate at 1066 speeds. At the current lower voltage it will only operate at DDR2 800 speeds. That could be related to my problem. I don’t know.

I’m not gonna bash AMD over this problem. The Spider platform seems very promising. Even with just two Radeon 3850s working coupled with a stock speed Phenom 9600 I am capable of playing a lot of games at 2560 * 1600 on my HP 3065 30″ monitor. I’m justed waiting for Catalyst 7.12 to enable the other two 3850s.

 
Comment by Michael Fox
2007-12-25 10:10:44

heh… “paper me”. that was classic! Maybe your next spoof of them should be “fix me” (”smarter choice” just isn’t doing it for me)

 
Comment by TLB errata???? wtf? Subscribed to comments via email
2008-01-04 00:10:25

I really don’t see why everyone is FUDing this processor (Fear, Uncertainty, Doubt). Its like every review site got paid by Intel to give bad reviews of AMD’s products. I personally own a 9500 and I can honestly say this: EVERY review site makes TOO BIG of a deal about this erratum. It NEVER locked up on me ONCE due to erratum. I never had corrupted DATA due to TLB erratum. I even OVERCLOCKED my processor to 2,640 mhz! Why in the world is everyone else acting like this processor is bad? I’ve done it on several processors and boards too… Evidently people don’t know how to properly update bioses or set bios settings or SOMETHING. Either that or you guys are buying whack-ass motherboards that suck. If there IS TLB erratum, I applaud AMD for being honest and announcing it, but my God people, Intel releases chips with erratum and doesn’t even make public statements of it. They delayed some off on the Penryn due to TLB Erratum!! My God! Everyone acts like its something that will destroy your data or eat your hard drive but let me tell you this as an owner of this product: It won’t.
Honestly. Don’t trust FUD websites.
Try it yourself, if you don’t like it, send it back and buy something else.

 
Comment by Jay
2008-01-15 22:26:48

Im building an AMD Phenom 9500 System right now. And I do intend to run some VM.

The patch can be bypassed with the OverDrive Utility which may not have 64 bit signing yet so F8 while booting and disable driver signing if your running Vista 64.

AMD applied the patch for those extreme “one off” times where a computer could lock up. But they give you the tools to undo that.

BTW Robert. you cant really overclock because the BlackBox edition CPUs have their multipliers unlocked (for overclocking) and the regular ones do not. Like Intel does.

 
Comment by Dom Subscribed to comments via email
2008-03-06 17:12:19

Personal Experience for me…. i own a
9500 also running at 2.6 Mhz!! with 2x 3850’s, GA-MA790fx-dq6 motherboard, bios has enable/disable options for TLB. Have tryd both options and never had a system crash?

Only had one problem - if you are buying this motherboard (GA-MA790fx-dq6) the Ram slots are very very close to the cpu and only a few upgrade cpu fan/heatsinks will fit! -_-

On the whole its a great gaming system!!

 
Comment by james Subscribed to comments via email
2008-10-03 03:26:42

I have a amd phenom 9550 quad core, where do i get the fix??? btw it is b2

 
Name (required)
E-mail (required - never shown publicly)
URI
Subscribe to comments via email
Your Comment (smaller size | larger size)
You may use <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong> in your comment.