Next-gen Intel chip benchmarks out already
- Thread starter clarkey01
- Start date
Those benchmarks don't bode well for Intel. Yes, a dual dual core Sossaman is faster than a X2 4800 but they conspicously left out a dual dual 275 AMD benchmark score. A dual Yonah at 2.0 loses against the dualcore AMD X2 3800+ that is already available? What the heck? Lower wattage won't mean a hoot to the desktop user and AMD already has low power Turions coming to compete in the notebook world.
StrangerGuy
Diamond Member
- May 9, 2004
- 8,443
- 124
- 106
The 1.5GHz DC Yonah doesn't look too good against the X2 3800+... Yeah I know it's 500MHz slower, but the performance gap is too big for comfort.
Rumor, speculation, crap.
They were extrapolating performance numbers.
Until we can see some real results on final silicon I'll just sit back and wait.
They were extrapolating performance numbers.
Until we can see some real results on final silicon I'll just sit back and wait.
If you didn't notice, the Sossman benches are for a dual processor dual core system. Plus the 2.0GHz score was extraploated off of the 1.5GHz score. If you give Intel the benefit of the doubt and figure that adding the second Sossman isn't going to give you exactly double the performance of a single Sossman in the benches then the probably scores fall into around the 500-550 range for a single Sossman, which puts it about on par with the X2 3800+ in clock for clock performance.
Of course if Cinebench performance scales more linearly with the addition of extra processors, or the performance for the Sossman doesn't scale exactly as they extrapolate at 2GHz, the you could actually see a Cinebench score for a single Sossman fall as low as 450 or so, which would put its performance lower than the X2 3800+ at 2Ghz.
Originally posted by: Elcs
I know this isnt an exact science but rather than extrapolating 1.5ghz up to 2.0ghz, why not underclock the 2.0ghz to 1.5ghz, ie. downclock the X2 to match the Intel job's ghz.
Or am I speaking too much sense?
I agree, that would make sense.
It looks like clock for clock, the intel chip has better performance.
JMO
Originally posted by: Elcs
I know this isnt an exact science but rather than extrapolating 1.5ghz up to 2.0ghz, why not underclock the 2.0ghz to 1.5ghz, ie. downclock the X2 to match the Intel job's ghz.
Or am I speaking too much sense?
First off, clock for clock comparisons are stupid because both architectures have different clockspeed targets. Underclocking the 3800 in order to match the Yonah makes no sense because there is no 1.5GHz SKU for the X2s and the family will probably top out at 2.8GHz, while Yonah will very likely have a 1.5GHz SKU. Also, Yonah will probably top out at around 2.33GHz (maybe 2.50, but I doubt it).
Originally posted by: fatty4ksu
Originally posted by: Elcs
I know this isnt an exact science but rather than extrapolating 1.5ghz up to 2.0ghz, why not underclock the 2.0ghz to 1.5ghz, ie. downclock the X2 to match the Intel job's ghz.
Or am I speaking too much sense?
I agree, that would make sense.
It looks like clock for clock, the intel chip has better performance.
JMO
Technically speaking it shouldnt. The only reason it does is cache. Yonah has a 14 stage pipeline and Hammer cores have a 12 stage
Originally posted by: fatty4ksu
Originally posted by: Elcs
I know this isnt an exact science but rather than extrapolating 1.5ghz up to 2.0ghz, why not underclock the 2.0ghz to 1.5ghz, ie. downclock the X2 to match the Intel job's ghz.
Or am I speaking too much sense?
I agree, that would make sense.
It looks like clock for clock, the intel chip has better performance.
JMO
Yep well the K8 is a 1999 design..About time.
Originally posted by: fatty4ksu
Originally posted by: Elcs
I know this isnt an exact science but rather than extrapolating 1.5ghz up to 2.0ghz, why not underclock the 2.0ghz to 1.5ghz, ie. downclock the X2 to match the Intel job's ghz.
Or am I speaking too much sense?
I agree, that would make sense.
It looks like clock for clock, the intel chip has better performance.
JMO
I don't know why I waste my time but...
In this benchmark, it looks like clock for clock, the intel chip has worst performance.
Notice, higher is better.
1º Simulated Yonah 2.0ghz: 521
BTW, optimistic score calculated if you ask me... taking as reference Yonah 1.5ghz score, I would say that 507-508 is more accurate for Yonah at 2.0ghz.
Never a CPU scales linearly... not to mention that 521 (from 383) is higher than a linear scale from 1.5ghz to 2.0ghz (We have: 1.5ghz, 383; so, 521 for 2.0ghz is a 36.03% higher score, when from 1.5ghz to 2.0ghz there is a 33,33% clock increase);
2º X2 3800+ 2.0ghz, 538
.
Originally posted by: PetNorth
Originally posted by: fatty4ksu
Originally posted by: Elcs
I know this isnt an exact science but rather than extrapolating 1.5ghz up to 2.0ghz, why not underclock the 2.0ghz to 1.5ghz, ie. downclock the X2 to match the Intel job's ghz.
Or am I speaking too much sense?
I agree, that would make sense.
It looks like clock for clock, the intel chip has better performance.
JMO
I don't know why I waste my time but...
In this benchmark, it looks like clock for clock, the intel chip has worst performance.
Notice, higher is better.
1º Simulated Yonah 2.0ghz: 521
BTW, optimistic score calculated if you ask me... taking as reference Yonah 1.5ghz score, I would say that 507-508 is more accurate for Yonah at 2.0ghz.
Never a CPU scales linearly... not to mention that 521 (from 383) is higher than a linear scale from 1.5ghz to 2.0ghz (We have: 1.5ghz, 383; so, 521 for 2.0ghz is a 36.03% higher score, when from 1.5ghz to 2.0ghz there is a 33,33% clock increase);
2º X2 3800+ 2.0ghz, 538
.
Sometimes you gotta laugh at how hard ppl try to conclude Intel have the better chip when they simply don't.
First, you are comparing a chip that is not even available with AMD chips that are widely available now. Second, the X2 3800+ (the lowest of AMD's x2 offering I might add) beats single dual core Yohan at stock speed (1.5Ghz) or "extrapolated speed" at 2ghz. The only way Yohan beats AMD X2 is when you run 2 Yohan. But come on, don't you think you should bench against 2x dual core opteron when you run 2 Yohan?
But if people find happiness in this kind of stuff, I guess it is their life.
A HIGHEND notebook processor can almost match a LOWEND desktop processor. Dont try to turn this benchmark into something positive for Intel, because it isnt. The only thing close to positive would be the nice scaling between the single cores and the duals, but that's it.
Dadofamunky
Platinum Member
- Jan 4, 2005
- 2,184
- 0
- 0
Funny how they had to use a total of four cores to allegedly beat a single 4800+. And of course they ONLY used CineBench which Intel is historically good at. Benchmarks can mean whatever you want them to mean, and most of the time they're horse manure to begin with. I should know - I wrote the book.
Author, The Benchmark Book (Prentice Hall, 1995)
Author, The Benchmark Book (Prentice Hall, 1995)
Originally posted by: fatty4ksu
Originally posted by: Elcs
I know this isnt an exact science but rather than extrapolating 1.5ghz up to 2.0ghz, why not underclock the 2.0ghz to 1.5ghz, ie. downclock the X2 to match the Intel job's ghz.
Or am I speaking too much sense?
I agree, that would make sense.
It looks like clock for clock, the intel chip has better performance.
JMO
Right, it?s a good thing you can buy one. And your observations (or delusions rather) are NOT opinions.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 24K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 19K
-
-
-
Facebook
Twitter