Athlons Move to 0.13 - clock increse as Tualatin?

[IIB]

should the athlon move to 0.13 process give it about the same boast as the Tualatin? i heard that the Tualatin can go almost up to 2GHZ if intell wanted it to...
so can we excpect athon with: 0.13(Thourghbird)+SOI(Barton)+Redisign(such as they did with palomino) to go up to 2.7+GHZ...????

if so i dont see any problem for the athlon
to stay competitve with the northwood... in terms of preformace that is...

oh... and sould we excpect SSE2 FPU in the athlons to come?
it will be a good move by AMD i think....

im no expert at cpu design or pc analysis but i thought ppl in the fourm might have some answers for me.
thank you.

 

[zengeos]

I've not heard of the Tualatin being able to go to 2ghz if Intel wanted. Since they couldn't get the P3 above 1ghz, it seems fair to expect up to 30 or 40% increase in performance at best. Still, that's a nice boost. I don't expect Athlon in Thoroughbred to go much above 1.9ghz. It may reach 2ghz, but I wouldn't expect it.

As for SSE2 on Athlon...I have yet to see it on any AMD roadmaps. I see it on the new Hammer chips, but not on the Athlon. That's not to say it won't happen, but AMD was pretty good about mapping the SSE support into Palomino and I would expect them to provide reasonably good notice for SSE2 on a future Athlon chip if it was planned.

Why do I only expect to see Thoroughbred to achieve around 2ghz clockspeed?
Because Palomino already produces portions of the Palomino in .13 traces; this according to Intel.

Mark-

When all else fails, throw your computer out the window!!!

 

[IIB]

about Tualatin going up to 2ghz... well i base this on:
my Thermaltake VOLCANO 6 HSF box says:
"Application:
AMD athlon 1.5GHz+
Intel Pentium3 1.Ghz &
TUALATIN up to 2.0GHZ"...
so i thought the guys at thermaltake know what there
talking about and the TUALATIN can go up to 2GHZ...

2Ghz (XP 2500+) for the athlon until 2qr 2002 sounds good to me dont think they need much more then that since the northwood by then shouldnt be above 2.5GHZ...
and barton is due in 2qr 2002 maybe the new SOI process and maybe some redisign will give the athlon a litle push at least until Hammer comes out...
i hope so coz even though i use intel competiton is good for all of us...

 

[jclw]

Over the years I've noticed that every ~0.6 micron decrease in die size results in an increase in clock speed of ~75%. The Katmai cores topped out at about 550MHz. The Coppermine cores go up to about 1GHz. The Tualatin 1.2 cores will do 1.6+ overclocked, and if intel released another stepping or two they should work there way up to about 1.73GHz (13x multiplier).

Following this logic the Thoroughbreds should hit ~2.7GHz (as the Palomino tops out at around 1.6GHz). I agree with you, I think the Thoroughbreds will be competive with the Northwoods.

- JW

 

[peteb]

[sarcasm]I often base my CPU preditions on the advertisment claims of no-affiliated heatsink companies too[/sarcasm]



-* <font color=red> Under Offer </font color=red> *-
email for application details

 

[AMD_Man]

I don't think the Palamino will top out at 1.6GHz. I think it may go up to at least 1.73GHz.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor

 

[peteb]

since it is only just released and is shipping 1.67 Jan 7th I think that is a safe assesment.

-* <font color=red> Under Offer </font color=red> *-
email for application details

 

[zengeos]

According to AMD's roadmap, I think Barton isn't due until sometime in 2H2002, not 2Q2002. AMD seems to be working on a 6 month development cycle. That makes sense since a shorter cycle would not be cost effective I think. (It takes several weeks I think, for a chip to go through he production process from start to finish) Add to that the possible confusion in products a shorter cycle would add. As it is, it looks like Palomino may have right around 6-9 month cycle, considering the 3-5 month rollout of the new chips.


When all else fails, throw your computer out the window!!!

 

[Guest]

Anyone noticed the trend of the process technology? The reduction from 0.6 to 0.35 to 0.25 to 0.18 then to 0.13 the margin of reduction is getting smaller and smaller? So the performance gains with be so marginal that u cant get anything out at the "real world".

And for that "hammer" SSE2 question, SSE2 will be only impletemented on the x87 FPU only. The ALUs that process integers will NOT have anything related to SSE2.

Play CS at only 41fps please...
that awp cant hit at 99fps...

 

[Matisaro]

I've not heard of the Tualatin being able to go to 2ghz if Intel wanted. Since they couldn't get the P3 above 1ghz, it seems fair to expect up to 30 or 40% increase in performance at best. Still, that's a nice boost. I don't expect Athlon in Thoroughbred to go much above 1.9ghz. It may reach 2ghz, but I wouldn't expect it.

Not really, the 1600mhz athlonxp is running about as hot as a 1.33ghz tbird, they released a 1.4ghz tbird, so thats another 133mhz for the axp, till they hit where the tbird was, however add to this that most tbirds could do 1600 with air cooling, and you can add a conservative 200mhz on to the axp. So were at 1.9ghz for the axp realistically, now reduce the heat/power 40% and you can easily score 2.6ghz or more. But its a moot point since hammer will be out before then.

Why do I only expect to see Thoroughbred to achieve around 2ghz clockspeed?
Because Palomino already produces portions of the Palomino in .13 traces; this according to Intel.

No, the traces are .18 microns, their gate widths are VERY tight for .18 micron, a testament to amds good fabbing, however this should not affect performance enhancement much from the die shrink, deffinatly not as much as intel claims, perhaps 5% lost, big deal. Since again, the hammer will be out before they hit the ceiling on the thoroughbred.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!

 

[lhgpoobaa]

ummm
of course!

thats silly!

u cant kep it linear

0.6 -> 0.35
0.35 -> 0.10
0.10 -> negative???


but we are seeing an exponential growth in mhz... so it must be working!


- Know thyself
- Seek
- Love
- Cherish Life
- And do not Yield!

 

[somerandomguy]

You have to look at the difference in percentages.
.18 is 72% of the size of .25
.13 is 72.2% of the size of .18 etc.

"Ignorance is bliss, but I tend to get screwed over."