View Full Version : Tualatin cpu beats athlon at same fsb.
Here is a interesting read thou most of its not in english.
http://www.overclocker.com.tw/article/0106/readgoodarticle.asp?banner=1&id=360
Froggy1
07-07-01, 12:48 PM
There is almost no english in that article, lol quick read :). One thing I did see was I think it said that the tualatin had 512KB of L2 cache. That would make a noticeable difference in speed no?
Now imagine how a couple of these would compare to these new dual Athlon systems. Cool, fast, and not necessarily requiring a nuclear reactor in your back yard to power it. Heck, you could probably get by fine with a generic 300W power supply for a dual Tualatin system and have plenty of juice to spare.
UnseenMenace
07-07-01, 05:56 PM
Froggy (Jul 07, 2001 12:48 p.m.):
There is almost no english in that article, lol quick read :). One thing I did see was I think it said that the tualatin had 512KB of L2 cache. That would make a noticeable difference in speed no?
If it is the case that this CPU indeed has 512 kb of L2 cache ( I couldnt read the review ), it could be the new Xenon type CPU and not the consumer release, if it is the Xenon version of the tualatin then perhaps the price should be compared before it is claimed that it beats it.
Yeah, so? Why would anybody want anything less? Those are precisely the ones I want. Besides the dual Athlon boards are targeted at the server market also. So, it's definitely a fair comparision. Furthermore with the dual Athlon boards costing what they do and the added cost of an expensive 650W enermax power supply the cost of the 2 platforms may not be that much different and it will probably be considerable cheaper that a dual P4 Xeon system.
Anyway, I don't think these will be refered as Xeons. I beleive they will use that name for the smp compatable P4s or Foster exclusively. I beleive these will be refered to a PIII-S. The "S" meaning that it's targeted at the server market. Anyway, here are some pics of the 1.26Ghz chip (http://www.watch.impress.co.jp/akiba/hotline/20010707/newitem.html).
Here is some guy running a Tualatin at 2ghz. aircooled.
http://www.clio.ne.jp/home/tatumiya/
Check out the fancy heatsink ???
How come they get all the chips 1st?
AmbientFiction
07-08-01, 01:59 PM
Placid (Jul 08, 2001 01:33 p.m.):
Here is some guy running a Tualatin at 2ghz. aircooled.
http://www.clio.ne.jp/home/tatumiya/
Check out the fancy heatsink ???
How come they get all the chips 1st?
Lmao is that a cup of noodles????
So what is that rapped around the heatsink??
AmbientFiction
07-09-01, 10:09 AM
bump
Looks like a plastic cup, maybe a big gulp :)
Froggy1
07-09-01, 02:03 PM
UnseenMenace (Jul 07, 2001 05:56 p.m.):
Froggy (Jul 07, 2001 12:48 p.m.):
There is almost no english in that article, lol quick read :). One thing I did see was I think it said that the tualatin had 512KB of L2 cache. That would make a noticeable difference in speed no?
If it is the case that this CPU indeed has 512 kb of L2 cache ( I couldnt read the review ), it could be the new Xenon type CPU and not the consumer release, if it is the Xenon version of the tualatin then perhaps the price should be compared before it is claimed that it beats it.
If you look below the graph that shows 3D Mark 2000 3D Results, it says in english Geforce 3 and Tualatin 512. I assumed the 512 remark meant 512kb L2 cache
Kingslayer
07-09-01, 02:09 PM
SP (Jul 07, 2001 04:45 p.m.):
Now imagine how a couple of these would compare to these new dual Athlon systems. Cool, fast, and not necessarily requiring a nuclear reactor in your back yard to power it. Heck, you could probably get by fine with a generic 300W power supply for a dual Tualatin system and have plenty of juice to spare.
I have seen some benchmarks of a dual AMD system and I wasn't the least bit impressed. The AMD guys seem to be making a big deal about them, but AMD needs to go back to the drawing board with them. They were slower in CPU benchmarks @1080 than my 800's @1002. And at least 100 points slower across the board.
Like I said, I'm not the least bit impressed with them.
asmodean
07-09-01, 02:16 PM
And here comes the AMD guy...
It has 512k of L2 cache. It's faster than a Thunderbird. It costs more than an equivalent Thunderbird. A tradeoff.
Not bad, though, as I've gotten used to AMD beating Intel hands tied.
If Intel chips weren't that expensive and unlockable, I'd propably be using an Intel chip myself.
*spazzed*
07-10-01, 02:04 AM
Placid (Jul 09, 2001 01:37 p.m.):
Looks like a plastic cup, maybe a big gulp :)
Whatever it is, it works well.....he he, it also says "BIG" on it :D
that's some crazy speed they got there ;D
I thought I saw the tulatin on intel's site at 512 cache?
**Yessir, that is correct (http://support.intel.com/support/processors/sspec/p3p.htm)!!
Anyway, cup-o-noodles cooling mod... don't try this at home? comes to mind
funnyperson1
07-10-01, 11:44 PM
well, at http://www6.tomshardware.com/cpu/01q2/010612/index.html...they have a tualitin 1.13..which they overclocked at 167 bs, and 1.4GHZ clock, and it came pretty darn close to the P4 and Athlon....and athlon got 200 bus, and P4..400...thats pretty damn nice
funnyperson1 (Jul 10, 2001 11:44 p.m.):
well, at http://www6.tomshardware.com/cpu/01q2/010612/index.html...they have a tualitin 1.13..which they overclocked at 167 bs, and 1.4GHZ clock, and it came pretty darn close to the P4 and Athlon....and athlon got 200 bus, and P4..400...thats pretty damn nice
The one they had at Tom's hardware was only a 256KB cache version. It was this ones little brother so to speak. The interesting one is the 512KB cache version or PIII-S. It's interesting because it may prove to be a better performing alternative to the dual AMD systems. A couple of those in a board likethis (http://www.viahardware.com/talkback/994878722,91371,.shtm) or the Iwill DVD266uR may be worthy of finally replacing my trusty old Tyan Tiger 100.
asmodean
07-11-01, 06:58 PM
Except Intel doesn't want you to use them in dual configs, I bet. It's really shaming for them to have P/// beat PIV...
Where can I find a Tualatin??
Only 1 place I know of now and they only sell it with a motherboard.
Here:
http://next.night.net/acb/showdetl.cfm?&DID=22&Product_ID=881430&CATID=2
Just have to wait for them to become more available.
asmodean (Jul 11, 2001 06:58 p.m.):
Except Intel doesn't want you to use them in dual configs, I bet. It's really shaming for them to have P/// beat PIV...
Where did you get this idea? The 512K PIII-S version supports SMP. It says so in the data sheet. It's designed for the server market where SMP is common. Intel isn't denying that these chips will work in a dual configuration. Quite the contrary they are smp certified by Intel. Intel obviously has no issues with these being used in a dual configuration. All of those articles you've read awhile back saying that Intel would somehow inhibit the use of these in dual configurations were written by people who didn't know what they were talking about and were just making it up. Take a look at the link to those motherboards in my post. Do you think Tyan, supermicro, and Iwill are building dual motherboards for chips that won't work in a dual configuration? I think not.
Also, as for the Tualatin beating the P4 that depends largely on the benchmarks. There are some benchmarks where nothing even comes close to the P4. The P4 is going to be a very fast chip. It's going to be the chip of the future. It's just that the future isn't quite here yet. The P4 is alot like the p6 family of processors when they were first introduced with the Pentium Pro. In the beginning the Pentium Pro was only of use in server applications and performed poorly on existing code. It was outperformed by the pentium MMX, but that soon changed. Right now the P4 doesn't appear to be all that fast using existing benchmarks, but with a few optimizations to the code compilers generate the P4 could gain significantly. I'm not talking about using SSE2 instructions either. I'm talking about simply optimizing code to reduce incorrect branch predictions. Anyway, a daul P4 Xeon system right now is fairly expensive. If it wasn't so expensive I might actually go that way as some of the software I run would probably benefit a good deal from it.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.