AMD with 81.5% of marketshare
- Thread starter KimB
- Start date
Pete said:
Athlon was generally faster than P3.
Athlon was faster than final P3s and initial P4s (Willamette).
Athlon XP traded blows with Northwood until the P4C, which gave Intel a few months of undisputed dominance until the late Athlon 64s arrived. Kind of interesting that the Athlon 64 and the Prescott were the only two new cores both companies launched around the same time and the Athlon 64 basically crushed the Prescott when the two should have been on about the same level. (mainly due to the Prescott not meeting design expectations; Willamette was a rushed product, and when Northwood took the lead the Athlon XP was already long overdue for retirement)
-The Pentium 3 Coppermine was more or less on par with the Thunderbird.
-The Tualatin was faster than the Thunderbird.
-Athlon XP was faster than the Willamette
-P4B was generally faster than the Athlon XP
-P4C was noticably faster than the Athlon XP
-Athlon 64 is on par with or slightly faster than the P4C and the Prescott excluding gaming, but also runs much cooler compared to the Prescott.
-Pentium M is faster and draws less power than the Turion.
-Core Duo is alot better than the Turion.
-Conroe/Merom will be alot faster than the Athlon 64.
You do now.
-The Tualatin was faster than the Thunderbird.
-Athlon XP was faster than the Willamette
-P4B was generally faster than the Athlon XP
-P4C was noticably faster than the Athlon XP
-Athlon 64 is on par with or slightly faster than the P4C and the Prescott excluding gaming, but also runs much cooler compared to the Prescott.
-Pentium M is faster and draws less power than the Turion.
-Core Duo is alot better than the Turion.
-Conroe/Merom will be alot faster than the Athlon 64.
xxx said:
You do now.
To generalize...ANova said:
If the app is singletheaded the A64 should be faster besides media encoding and some other odd apps that like netburst.
If the app is multithreaded than at worst the P4C and prescott should be on par and best eat it for lunch... usally the latter.
edit- theres also the socket change amd did... I'm referring to S939 A64s.
and the conroe will be alot faster than the dual core p4s also
I'm also read about how availability will be an issue at first ...
Back to topic.
P4B and up were good compared to the athlon64 in apps that are SMP aware because of their SMT capabilities, without it they aren't that competative.
I also have an issue with how they are compared.. to me comparing model numbers and ghz shouldn't be used, instead the MSRP (or streetprice) should be used since I think that's how intellgent people buy things, for joe consumer it's fine tho, since he don't know any better.
But for people in the know they shouldn't just go"hmm.. I can get a 3200+ for XXX or a P4C 3.2ghz for XXX.." they should look at bang for your buck and that's why amd has a huge marketshare excluding oems I think.
Last edited by a moderator:
ANova said:
Price maybe should be a consideration...
Max clock speed definetely should be, Coppermine and Tualatin ultimately lost due to lower clock speed than the Athlons, even if they could keep up on a clock for clock basis. Not to mention DDR memory gave the Athlons a huge advantage in some things.
Depends on price. Conroe may immediately get some of the enthusiast marketshare back in Intel's court (depending upon performance, and to some extent price), but I doubt it'll have much effect on the retail marketshare unless the costs are low, and Intel ships Conroe versions of the Celeron.Tahir2 said:
It doesn't really matter that the Conroe isn't out yet because what is known is that it performs extremely well, will be out in a matter of a few months and the only update for the Athlon 64 is a new socket with DDR2 support until at least next year. Price is a non factor as well, the high end Conroe (E6700) will retail for about $530 while the next down will go for a mere $316 and these things overclock like crazy. So for about $209 you will be able to get noticably better performance than an FX-60.Tahir2 said:
Last edited by a moderator:
Well an air cooled Conroe at about 2.6GHz is about a 6GHz P4 LN2 cooled it seems, most conroe's can go past 3Ghz on air as well and be around 3Ghz with default voltages.
AMD will have to drop it's prices and it will still be slower. looks like AMD is in for a tough time for the next 6 months or so.
Conroe will be at the high end and cheap dual P4's will be at the low end so some realignment of AMD prices is in order if they want to continue to sell in both market segments I think.
AMD will have to drop it's prices and it will still be slower. looks like AMD is in for a tough time for the next 6 months or so.
Conroe will be at the high end and cheap dual P4's will be at the low end so some realignment of AMD prices is in order if they want to continue to sell in both market segments I think.
chavvdarrr
Veteran
Where? in "super-ultra-pi" ?dizietsma said:
afaik, 2.66 Conroe with 4Mb L2 cache seems as fast as 3.0-3.2 k8
chavvdarrr said:
It's producing better results than that (overclocked) speed of K8 at it's stock speeds in 3dmark now I am afraid .... at 3Ghz it's also 20% faster than a 3Ghz FX60 in pifast which has traditionally favoured AMD.
The problem is that a lot of people who are fond of AMD will have a lot of trouble accepting that the AMD K8 is no longer king of the hill when it comes to desktop cpu's.
dizietsma said:
Most conroes can go past 3ghz on air? Considering they're still on production samples, I wouldn't make any guarantees about that. People were saying most Pentium M's could hit near 3ghz, but I've heard of quite a few which cap out at 2.2ghz. Just because there's been some promising early results doesn't mean that anywhere near the majority will clock that high.
Depending on the app, a 3-3.2ghz K8 would match a 6ghz p4.
3dmark has always favored Intel for the 3dmark score, I'd imagine they're able to eek slightly more performance out of their AGP/PCI-Express buses than AMD mobos do.
And it will be interesting to see how AMD responds. Right now it's Intel who's cutting their prices to rock bottom, AMD is standing still. Too bad prescotts run so hot and that a maxed out overclock on a cheapie $130 Intel dual core still couldn't touch an overclocked 3800+.
Tahir2 said:
ANova said:
It really does matter Conroe is not out yet. You cannot buy it yet, and if that is the miracle product for Intel clawing back some of its cred on the retail side (as we are discussing here) then it helps if you can buy the thing.
I already stated Conroe is likely to win Intel some marketshare back when it is launched.. and therefore was there a need to only quote half of my post?
I will add to further this debate - it takes years to build momentum for a market shift, it never happens quickly and by that time if AMD responds (and being a smaller company it probably can respond in a timely fashion) Intel will need another trick up its sleeve.
One important thing to consider is that Intel is currently on 65nm while AMD is still on 90nm. What should be more surprising than Conroe being faster than AMD's current high-end A64's is that Intel's current 65nm CPU's are slower.
It seems conceivable to me that AMD could well put things on par with Intel simply by migrating their existing design to 65nm. I'm not currently aware of when AMD plans to do this, however.
It seems conceivable to me that AMD could well put things on par with Intel simply by migrating their existing design to 65nm. I'm not currently aware of when AMD plans to do this, however.
One other little issue among the enthusiast market : Intel is making a big push to get VT into places where it's needed. i.e. enabling Xen to virtualise without kernel changes needed in guest OS'es, enabling Intel Macs to run Windows XP in a VM with tools like Parallel Workstation.
Yes, AMD's Pacifica is coming, and looks to be perhaps a little better than the first generation Intel VT, but not only is Intel 6 months ahead in terms of implementation, they also seem to be investing heavily in their take on virtualization extensions and delivering to all parties busy building around CPU virtualization extensions.
I'm not really sure how much of a niche this area is, but to be honest at the moment it's quite tricky to decide which server CPUs to invest in given that Xeons are already VT+EMT64 and Opterons, while definitely faster, lack Pacifica. Conroe will be a relief on that front since it'll remove any concerns about proper Pacifica support in future Virtualization hypervisors over the 3 year life of a server.
Yes, AMD's Pacifica is coming, and looks to be perhaps a little better than the first generation Intel VT, but not only is Intel 6 months ahead in terms of implementation, they also seem to be investing heavily in their take on virtualization extensions and delivering to all parties busy building around CPU virtualization extensions.
I'm not really sure how much of a niche this area is, but to be honest at the moment it's quite tricky to decide which server CPUs to invest in given that Xeons are already VT+EMT64 and Opterons, while definitely faster, lack Pacifica. Conroe will be a relief on that front since it'll remove any concerns about proper Pacifica support in future Virtualization hypervisors over the 3 year life of a server.
What makes you think that AMD will not only be able to match the 30%+ advantage the Conroe has over the A64 but increase that enough for 'Intel to need another trick up its sleeve'? Sounds like wishful thinking to me.Tahir2 said:
Not necessarily. The Prescott was the first chip to be on the 90nm process and what did that get it? The Conroe is much more than just a die shrink, the A64 may be able to close the gap by moving to 65nm but it'll need more than that to beat it. Though, Intel will be adding the memory controller onto its chips in addition to cache increases by late next year as well. Oh and AMD will be moving to 65nm Q1 of next year.
chavvdarrr
Veteran
AFAIK Pacifica comes with AM2 Opterons.skazz said:
My company had bought few servers last month, 2 of them supposed to run several VMs (through VMware). First with Xeons, the second with Opterons. (the second machine was bought a week after first arrived - and below you can see why we bought Opterons)
The bad thing - the Xeons had no VT support (although they support AMD x64 , or if someone wants to believe Intel created their own x86 64-bit extension - EMT64
).Few quick questions to our suppliers revealed that supply of VT-enabled CPUs is limited, and we decided not to change CPUs right now and go get K8 based server in order to compare products with our own hands... we'll buy more machines
2nd bad thing - the 2 systems are identical as periferals (scsi hdds, etc), but Xeons run slower under 64bit 2003 R2. It can be seen easily when firing same VMs on both machines. And we need 64-bit OS, because wanna use all 8GB of RAM (and no we don't wanna buy AS or DC). Also running some tasks alocating CPUs via affinity revealed unsatisfying scaling after 2nd core.
3rd bad thing - VMWARE does not support running 64-bit clients on 64-bit host on Intel CPU without VT - even on such cpus, support is "experimental", in same time they support running same scenario on ANY AMD K8 CPU rev E or later (or it was D? and C for Opterons? hmm, any AMD CPu made since mid-2005). Even more, they support running 64-bit client on 32-bit host (not that we will try this
). And by explanation why it seems really resonable (after rev.C K8 supports memory segmenation, this is supported only on Intel CPUs with VT)So, tell me now - who is shipping CPUs better suited for running VMs? Here and now I know the answer, Maybe after few months when Conroe and Woodcrest come (and their 64-bit performance is OK... because I haven't seen any 64-bit tests with Conroe), and few months later when there will be rev.1.1 motherboards for them the answer will different. Who knows. Competition is good
Last edited by a moderator:
chavvdarrr said:AFAIK Pacifica comes with AM2 Opterons.
.....And we need 64-bit OS, because wanna use all 8GB of RAM (and no we don't wanna buy AS or DC). Also running some tasks alocating CPUs via affinity revealed unsatisfying scaling after 2nd core.
3rd bad thing - VMWARE does not support running 64-bit clients on 64-bit host on Intel CPU without VT - even on such cpus, support is "experimental", in same time they support running same scenario on ANY AMD K8 CPU rev E or later (or it was D? and C for Opterons? hmm, any AMD CPu made since mid-2005). Even more, they support running 64-bit client on 32-bit host (not that we will try this
So, tell me now - who is shipping CPUs better suited for running VMs? Here and now I know the answer, Maybe after few months when Conroe and Woodcrest come (and their 64-bit performance is OK... because I haven't seen any 64-bit tests with Conroe), and few months later when there will be rev.1.1 motherboards for them the answer will different. Who knows. Competition is good
I view a server as something which will be running for 3 years, or 4 in some cases. In 3 years time the state of the art in terms of virtualization software will have moved on dramatically (e.g. minikernel hypervisors relying on next wave CPU virtualization extensions), so considering a certain amount of "future proofing" is no bad thing.
For now, if you're running VMWare Server on a Windows host, then AMD CPUs are definitely the best choice, as long as you are not likely to want to change the function of that machine any time soon. However, Xen gains significantly from Intel VT already, and the upcoming Windows Virtualization product (codename Viridian) is also going to be hooking into CPU virtualization extensions. So thinking about the longer life cycle of the server, not having those extensions may well force "early retirement" of the hardware (or at least moving it to a different function). Difficult to predict, and very different to typical server functions.
On a side note (heading swiftly offtopic): VMWare Workstation 5.5 and VMWare Server 1.0 both run VMs in 32-bit mode, even within a x64 host, using WOW64. So it really doesn't matter whether you run 64-bit guests inside a 32-bit or 64-bit host from the point of view of functionality. Of course, for host memory management you definitely benefit from x64 host OS as compared to running a 32-bit host OS in /PAE mode.
Regarding host OS costs : Did you notice the announcement from Microsoft a month or so ago where they said that buying W2K3 Enterprise Edition gives one rights to 4 additional instances of that OS for free inside virtual machines on the same host? (not specifically virtual server either). Oh, and their other announcement that you only have to pay Windows OS license fees for running VMs, not for every single VM sitting on a hdd inactive. (I guess this one is only for enterprise customers though).
cheers, Skazz
Similar threads
