The AMD Execution Thread [2007 - 2017]

Status
Not open for further replies.
Except that neither SIS nor VIA have shown an ability in recent years to produce a stable, performant chipset with decent and competitive integrated graphics. All of which are required if one wishes to take on Intel in the OEM market.

The only options available at the time would have been Nvidia and ATI. Ignoring the fact that Nvidia would have been a far more expensive purchase, ATI chipsets were at least as stable as Nvidia chipsets (some would argue more stable) with overall much lower power consumption.

For what AMD was after, ATI really was the only option available. They would be in far worse shape had they picked up either VIA or SIS.

I have a feeling that had Barcelona and Phenom performed as well as AMD was expecting, that everyone would be singing their praises for aquiring ATI even if it was only for the mainboard chipset expertise.

And lets not forget. Out of all of AMD's divisions ATI is the closest to going back into profitability. If they excute well on that front it isn't that far fetched to think that ATI might end up keeping AMD afloat long enough for them to be competitive in the CPU business again.

Will that happen? Who knows. Rv670 is promising. But is a sign of things to come and is it a sign that AMD at the helm of ATI will be able to keep them competive in the midrange while challenging for the high end?

Or will AMD hunker down with ATI and attempt to dominate the more lucrative mid-range hoping that superior/low cost mid-range parts will be able to offset the loss of a "halo" effect of dominant High End parts and the free advertising that brings?

After all prior to R300, that is how ATI managed to survive. By targeting the OEM market with low cost (production cost) low to mid range parts. Nvidia was more perfomant but ATI's chips were far far cheaper for OEMs.

Regards,
SB

Yes, yes, yes, I am quite acquainted with that particular argument. It's fine and dandy. I'm not arguing the fact that ATi was superior to either Via or Sis as a chipset maker/company whatever. But they were far too expensive especially due to that. Having a top-notch chipset maker means jack-squat when you're so far into debt that you're pretty much fucked. Even if Phenom had rocked the soxxorz off of everyone they would've been in fairly murky waters. And I certainly doubt AMD ever thought that Phenom would dominate in typical desktop loads. They were always aware of their lacking IPC when compared to Core 2, and they must've become aware that they can't clock the thing high enough.

Look at it differently:what you have now as a chipset is an AMD 790FX, correct?I theorise that for the average buyer, the one that actually cares about buying a platform like Spider because crap from the same vendor works better together or whatever doesn't really care if the above is the ATi RD700 in reality or if it were the Via PT-Gaga or some SiS thingie. The stability and performance delta, whilst certainly noticeable for enthusiasts, is less aparent for your average dude toying with Excel and some games, with no OC and so on. And Via is actually fairly decent, though not great(as a side-note, ATi has about 0 extraordinary chipsets done in its history, so I'm not sure that they qualify themselves in the great-chipset maker category),IMHO. Not enough to warrant the risk that the ATi aquisition posed, not to bet the wife and house on it.

Yeah, they probably wanted to compete with Intel in the future and took Intel's future direction into account...but it's quite silly to think that you can survive a no holds barred battle with Intel, and try competing with them through aggresive expensive aquisition, thus trying to match them in firepower. Being clever about it and waiting in the shadows, having a similar arsenal even though lower-powered, hunting for the opening that gives you the chance of sucker-punching would've been a better way to do things. Heck, it's what AMD did for most of its life...until they actually started believing their own marketing, and became cats who saw themselves as lions in the mirror.
 
HD 3800 has serious supply constraints

Despite what you read in the media

An add-in board partner has come on the record and said that "Despite what you have read and observed in the media, there are serious supply constraints [of the whole HD 3800 series]."

As we believe that these brave people can get in trouble with AMD's sale machine we will keep their name for ourselves, but this certainly implies that AMD still has an issue to ship enough of the cards to feed the hungry market.

The same source continues woth some strong words. "The shortage is so severe that we are unable to supply our channel and other vendors at this time."

Here goes the bubble that the HD 3870 is widely available while the 8800 GT is in severe short supply and it looks like you might have trouble to buy any of these two cards in 2007. DAAMIT, it had to happen at the time where 20+ hot games are out where I can personally recommend at least five of them, which all could do with a graphics card upgrade for a smooth frame rate.

http://www.fudzilla.com/index.php?option=com_content&task=view&id=4500&Itemid=34

I wonder if they represent the average company.
 
Look at it differently:what you have now as a chipset is an AMD 790FX, correct?I theorise that for the average buyer, the one that actually cares about buying a platform like Spider because crap from the same vendor works better together or whatever doesn't really care if the above is the ATi RD700 in reality or if it were the Via PT-Gaga or some SiS thingie. The stability and performance delta, whilst certainly noticeable for enthusiasts, is less aparent for your average dude toying with Excel and some games, with no OC and so on. And Via is actually fairly decent, though not great(as a side-note, ATi has about 0 extraordinary chipsets done in its history, so I'm not sure that they qualify themselves in the great-chipset maker category),IMHO. Not enough to warrant the risk that the ATi aquisition posed, not to bet the wife and house on it.

790FX isn't supposed to the "cash cow" if you will of chipsets. It's supposed to be a high visibility chipset marketed towards Enthusiants and Hardware review sites. It's aiming to try to compete for the enthusiast/gamer who tend to be quite loud and braggadocious. Free marketing if you will.

The cash cow of the chipset business is in stable chipsets with integrated video. Something to woo business customers and OEM customers. Neither VIA nor SIS have had anything I would consider even remotely appropriate for those situations. Just as an aside I have a ATI chipset MB running a server 24/7 that also used to serve as a gaming machine. Far more than I can say for any VIA, SIS or Nvidia consumer chip. I do have an Nvidia based Sun Workstation, of course it's NOT a reference board and maybe that's why it's actualy stable for 24/7 use + gaming as opposed to almost all consumer Nvidia boards I've used, which are generally geared towards performance first and stability second.

The point of the above isn't to rag on Nvidia. But to show that super dooper high performance overclocked balls to the walls boards aren't what sells in volume. It's the ability to have a stable OEM/Business integrated platform. And that's just what ATI brought.

Yeah, they probably wanted to compete with Intel in the future and took Intel's future direction into account...but it's quite silly to think that you can survive a no holds barred battle with Intel, and try competing with them through aggresive expensive aquisition, thus trying to match them in firepower. Being clever about it and waiting in the shadows, having a similar arsenal even though lower-powered, hunting for the opening that gives you the chance of sucker-punching would've been a better way to do things. Heck, it's what AMD did for most of its life...until they actually started believing their own marketing, and became cats who saw themselves as lions in the mirror.

I don't believe it was a mistake or that AMD couldn't go head to head with Intel. I think where the major problem came in is that AMD in the form of Hector Ruiz had an overly dim view of Intels competitiveness. Ruiz had the feeling that with A64 and the breakthrough they had with that, that they could no do no wrong. And that with the technological superiority they could stay ahead of Intel as it continued to "blunder" along. Ruiz I think never really gave serious thought that even though Intel may have been technologically behind (NUMA, On die memory controller, etc), that they could still excecute well and quickly.

Also, that Intels resources meant that even if they lagged (IE - during the P4 years), that their ability to produce chips and quickly ramp up production of chips meant that AMD would, at least for some time to come, remain production constrained and thus even with high demand, AMD couldn't make much of a dent in Intel even if there was extraordinary demand for AMD chips due to low production volumes (in comparison to Intel).

All things which mean that AMD couldn't afford to assume that Intel couldn't compete, something Ruiz just didn't seem to get. Granted, you need to have some sort of optimism even in the face of adversity, but Ruiz had so much that it blinded him to just how dangerous Intel was. And it blinded him to the fact that as with all tech companies, you will not always have success in everything you design.

In other words too much patting yourself on the back leaves you wide open to be blind sided by the opposition. Happened to Intel and is now happening to AMD.

Regards,
SB
 
DAAMIT, it had to happen at the time where 20+ hot games are out where I can personally recommend at least five of them, which all could do with a graphics card upgrade for a smooth frame rate.

The 8800 GTX / Ultra are in abundance. You might even be able to find them fairly cheap on eBay or other For Sale forums. Of course it wont be quite the bargain that a 8800 GT or HD 3870 is, but it's available and incredibly fast.
 
Of course it wont be quite the bargain that a 8800 GT or HD 3870 is, but it's available and incredibly fast.

I guess the author was focusing on the new product lines. I posted here because it shows the lack of hope for future nice products and an atmosphere of despair.
 
790FX isn't supposed to the "cash cow" if you will of chipsets. It's supposed to be a high visibility chipset marketed towards Enthusiants and Hardware review sites. It's aiming to try to compete for the enthusiast/gamer who tend to be quite loud and braggadocious. Free marketing if you will.

The cash cow of the chipset business is in stable chipsets with integrated video. Something to woo business customers and OEM customers. Neither VIA nor SIS have had anything I would consider even remotely appropriate for those situations. Just as an aside I have a ATI chipset MB running a server 24/7 that also used to serve as a gaming machine. Far more than I can say for any VIA, SIS or Nvidia consumer chip. I do have an Nvidia based Sun Workstation, of course it's NOT a reference board and maybe that's why it's actualy stable for 24/7 use + gaming as opposed to almost all consumer Nvidia boards I've used, which are generally geared towards performance first and stability second.

The point of the above isn't to rag on Nvidia. But to show that super dooper high performance overclocked balls to the walls boards aren't what sells in volume. It's the ability to have a stable OEM/Business integrated platform. And that's just what ATI brought.



I don't believe it was a mistake or that AMD couldn't go head to head with Intel. I think where the major problem came in is that AMD in the form of Hector Ruiz had an overly dim view of Intels competitiveness. Ruiz had the feeling that with A64 and the breakthrough they had with that, that they could no do no wrong. And that with the technological superiority they could stay ahead of Intel as it continued to "blunder" along. Ruiz I think never really gave serious thought that even though Intel may have been technologically behind (NUMA, On die memory controller, etc), that they could still excecute well and quickly.

Also, that Intels resources meant that even if they lagged (IE - during the P4 years), that their ability to produce chips and quickly ramp up production of chips meant that AMD would, at least for some time to come, remain production constrained and thus even with high demand, AMD couldn't make much of a dent in Intel even if there was extraordinary demand for AMD chips due to low production volumes (in comparison to Intel).

All things which mean that AMD couldn't afford to assume that Intel couldn't compete, something Ruiz just didn't seem to get. Granted, you need to have some sort of optimism even in the face of adversity, but Ruiz had so much that it blinded him to just how dangerous Intel was. And it blinded him to the fact that as with all tech companies, you will not always have success in everything you design.

In other words too much patting yourself on the back leaves you wide open to be blind sided by the opposition. Happened to Intel and is now happening to AMD.

Regards,
SB

I don't agree but the argument would be fairly pointless. Reality though doesn't quite agree with you, it being composed of AMD's financials, it's crash post ATi aquisition, ATi's "nothing-to-write-home-about" history when it comes to making chipsets, and a few other things. I'd really like to live in a world where billions in debt are just the consequence of one guy being too full of himself, and where a company that's about 20 times smaller than it's main competitor can go head to head with it on all fronts, simply because they're cool like that. It would definitely make my job easier...sadly, it's usually far more complex. Maybe it's just me though.
 
Chip problem limits supply of quad-core Opterons

From TechReport article:
...
An industry source at a tier-two reseller told The Tech Report that the TLB erratum has led to a "stop ship" order on all Barcelona Opterons. When asked for comment, spokesman Phil Hughes said AMD is shipping Barcelona Opterons now, but only for "specific customer deals." Industry sources have suggested to TR that those deals are high-volume situations involving supercomputing clusters. Such customers may run workloads less likely to be affected by any workarounds for the erratum that reduce L3 cache performance, and those customers could potentially consume hundreds of thousands of CPUs. Our sources indicate, and the current availability picture would seem to confirm, that quad-core Opterons are not shipping to OEMs or the channel more generally.

News of this problem is notable because it confirms that the TLB erratum affects Barcelona server processors as well as Phenom desktop CPUs, and that the problem impacts AMD's quad-core processors at lower clock speeds. AMD's initial public statements about the erratum and the delay of the 2.4GHz Phenom seemed to imply that the issue was closely related to clock frequencies. The Opteron 2300 lineup spans clock speeds from 1.7GHz to 2.0GHz. Those CPUs' north bridge clocks, which determine the frequency of the L3 cache, range from 1.4GHz to 1.8GHz.

The erratum is present in all AMD quad-core processors up to the current B2 revision. AMD has said a revision B3 is in the works and expected in Q1. One source told TR that large quantities of B3 chips might not be available until the end of Q1.
...
At present, Microsoft doesn't offer a Windows hotfix to address the problem, and our sources were doubtful about the prospects for such a patch. CPU makers have oftentimes addressed errata via updates to the processor's microcode, but such a fix for this problem also appears to be unlikely.

When it rains it pours... Bolding is mine, to indicate a likely reason why shortages exist or why it may be in your best interest to not purchase this CPU now.
 
Does it? Can indicate exactly the opposite.

The other cards cited by Brit are not in =<65nm nodes. That's why I am afraid of everything that comes lower, or in other words, all future GPUs. But I hope that this only happens to 65nm.

Did anyone try to overclock Phenon up to 4Ghz (liquid helium cooling perhaps) and see what happens?
 
From TechReport article:


When it rains it pours... Bolding is mine, to indicate a likely reason why shortages exist or why it may be in your best interest to not purchase this CPU now.

I was kind of wondering why this hadn't been disclosed sooner.
Phenom and Barcelona are after all the same chip, with only a minor stepping difference, if that.

It would be an interesting thing if it turned out Barcelona didn't have this problem.
 
The article also said that AMD disclosed that a likely bios fix would impose a 10% performance penalty, while another source pegged the performance penalty at 10 - 20%. I'm not sure if thats in specific cases or all cases. Whichever doesn't much matter; it's pretty damming to have the company performing so poorly all around (GPU, CPU, Chipset).
 
The article also said that AMD disclosed that a likely bios fix would impose a 10% performance penalty, while another source pegged the performance penalty at 10 - 20%. I'm not sure if thats in specific cases or all cases. Whichever doesn't much matter; it's pretty damming to have the company performing so poorly all around (GPU, CPU, Chipset).


IIRC, that was the initial BIOS fix in order to get the chips out the door, but a later BIOS is supposed to provide a proper microcode fix that will get the speed (such as it is) back to where it should be without crashing.
 
From TechReport article:


When it rains it pours... Bolding is mine, to indicate a likely reason why shortages exist or why it may be in your best interest to not purchase this CPU now.

It's good that the "hundreds of thousands" of Barcelona's AMD "shipped" this quarter were largely imaginary, since those kinds of recalls are much easier than the real ones.
 
It's good that the "hundreds of thousands" of Barcelona's AMD "shipped" this quarter were largely imaginary, since those kinds of recalls are much easier than the real ones.

It was probably a "bathtub" kind of ocean, so that's very likely... :LOL:
 
Personally, it sounds more like that the initial BIOS fix will drop performance 10% from where they are now, and the future microcode fix will only return it back to the original performance.

edit: A new Tech Report article seems to confirm this:

http://www.techreport.com/discussions.x/13724

Perhaps so, perhaps not. It seems like some reviews were using overclocked parts based on AMD's guidance. What they gain from the fix may not be that much, and could see it just enough to offset the loss. Tech Report states the following:

On a related note, AMD PR consistently denied TR's request for samples of the production Phenom 9500 and 9600 models in the days following the product launch, until we informed them that we'd ordered a CPU from Newegg. We received a production sample of the Phenom 9600 from AMD shortly thereafter, followed by the 9500 we purchased at Newegg.

We don't yet have a BIOS with the workaround to test, but we've already discovered that our Phenom review overstates the performance of the 2.3GHz Phenom. We tested at a 2.3GHz core clock with a 2.0GHz north bridge clock, because AMD told us those speeds were representative of the Phenom 9600. Our production samples of the Phenom 9500 and 9600, however, have north bridge clocks of 1.8GHz. We've already confirmed lower scores in some benchmarks.

Given everything we've learned in the past few days, our review clearly overstates Phenom 9600 performance, as do (more likely than not) other reviews of the product. We can't know entirely by how much, though, until we can test a Phenom system with the TLB erratum workaround applied.
 
Personally, it sounds more like that the initial BIOS fix will drop performance 10% from where they are now, and the future microcode fix will only return it back to the original performance.

edit: A new Tech Report article seems to confirm this:

http://www.techreport.com/discussions.x/13724

There is no future microcode update as per TechReport's latest piece-there's only the current one, that's implemented through a bios update, that degrades performance by 10%. They'll fix it in silicon with B3(they're hoping) and replace 9500 and 9600 with 9550 and 9650 to denote the new B3 revisioned chips. It's like the entire cosmos is having fun at AMDs expense, fubaring everything they try to do:)
 
Perhaps so, perhaps not. It seems like some reviews were using overclocked parts based on AMD's guidance. What they gain from the fix may not be that much, and could see it just enough to offset the loss. Tech Report states the following:

AMD is up shit creek without a paddle right now. WTF are they thinking?
 
AMD is up shit creek without a paddle right now. WTF are they thinking?

I think they were panicked by Intel. Core2 and now Penryn kicking AMDs arse, so instead of holding off on Phenom until they had a working stepping at the required speed, or RD790 until they had SB700, they just went ahead and launched what they had to get something out there. "Look at us, we still have products, they are cheap, they all work together."

Unfortunately it bit them on the ass, as stuff doesn't work right, and by hanging GPU/CPU/chipset all to each other, the good bits are dragged down by the bad bits. AMD went for a all-or-nothing, all-eggs-in-one-basket, and screwed it up.

Things would have been dramatically different if Phenom had arrived at 3 ghz. Not enough to beat Intel's Penryn, but enough to keep hanging onto Intel's coat-tails, and enough to make Spider look really attractive if pitched at the right price. Instead, AMD and their current offerings are simply giving the company a poor reputation as underperforming, overpriced, late, incapable of executing, buggy, etc. Once your reputation is screwed, so is the confidence in your stock price, and your ability to get/make money.
 
Status
Not open for further replies.
Back
Top