Ok just a simple question: how much faster will a SINGLE core of an core
i7 975 be vs an athlon XP 2600+ (2 Ghz) ? I suppose one get the core i7
with DDR3 and best mobo available for this cpu at this moment, and the
athlon is with DDR , NF7 mobo.
Do you guys have a precise idea? I can't find a benchmark with the
athlon XP 2600+ and the core i7 975...

blibloo wrote:
> Ok just a simple question: how much faster will a SINGLE core of an core
> i7 975 be vs an athlon XP 2600+ (2 Ghz) ? I suppose one get the core i7
> with DDR3 and best mobo available for this cpu at this moment, and the
> athlon is with DDR , NF7 mobo.
> Do you guys have a precise idea? I can't find a benchmark with the
> athlon XP 2600+ and the core i7 975...

http://www.cpubenchmark.net/cpu_list.php
Well you've selected the fastest cpu one the list to
compare with an obsolete one. 15+ times faster.

On 09/19/2009 04:45 PM, Pen wrote:
> blibloo wrote:
>> Ok just a simple question: how much faster will a SINGLE core of an
>> core i7 975 be vs an athlon XP 2600+ (2 Ghz) ? I suppose one get the
>> core i7 with DDR3 and best mobo available for this cpu at this moment,
>> and the athlon is with DDR , NF7 mobo.
>> Do you guys have a precise idea? I can't find a benchmark with the
>> athlon XP 2600+ and the core i7 975...
>
> http://www.cpubenchmark.net/cpu_list.php
> Well you've selected the fastest cpu one the list to
> compare with an obsolete one. 15+ times faster.
thanks for the link, i've already browse this, but it doesn't tell me
whether or not the cpu mark is obtained by using all the cores, i wanna
know how ONE core and one only performs vs the Athlon...
Would like to have a simple unithreaded benchmark like zipping a file or
encoding a wav file into mp3, which would tell me x seconds for the
Athlon and y for the i7 ...

blibloo wrote:
> On 09/19/2009 04:45 PM, Pen wrote:
>> blibloo wrote:
>>> Ok just a simple question: how much faster will a SINGLE core of an
>>> core i7 975 be vs an athlon XP 2600+ (2 Ghz) ? I suppose one get the
>>> core i7 with DDR3 and best mobo available for this cpu at this moment,
>>> and the athlon is with DDR , NF7 mobo.
>>> Do you guys have a precise idea? I can't find a benchmark with the
>>> athlon XP 2600+ and the core i7 975...
>>
>> http://www.cpubenchmark.net/cpu_list.php
>> Well you've selected the fastest cpu one the list to
>> compare with an obsolete one. 15+ times faster.

> thanks for the link, i've already browse this, but it doesn't tell me
> whether or not the cpu mark is obtained by using all the cores, i wanna
> know how ONE core and one only performs vs the Athlon...
> Would like to have a simple unithreaded benchmark like zipping a file or
> encoding a wav file into mp3, which would tell me x seconds for the
> Athlon and y for the i7 ...

The easy way to tell Passmark is comparing *all* cores, is to compare
the performance of a Q9650 3.0GHz Quad core to an E8400 3.0Ghz Dual core.
The quad is 4459. The dual is 2046. So the Passmark benchmark must
be multithreaded, and comparing all available cores combined.

SuperPI is single threaded. This is an Athlon XP 2600+ Thoroughbred @ 2150mhz
running SuperPI 32M. Execution time is 46min 59sec 940ms. (I just checked
a table for processors, and 2600+ is listed as 2083MHz, so close enough.)

http://hwbot.org/result.do?resultId=593780

Intel Core i7 Processor Extreme Edition I7-97 at 3.33GHz. Well,
it appears all the submissions for this processor, are overclocked.
So this is running at 3.855GHz. Execution time is 9min 56sec 420ms.

http://hwbot.org/result.do?resultId=850201

We need to scale that result a bit. Convert to seconds first.
9min 56sec 420ms = 596.42 sec.

3855
596.42 * ---- = 690 seconds (if run at stock speed)
3330

Now, convert the Athlon2600+ Thoroughbred time to seconds.
46min 59sec 940ms = 2819.94 seconds.

Therefore, a Core i7 975 single core is 2819.94/690 = roughly 4.08 times faster.

You can validate the SuperPI benchmark on your own computer.
Run this, selecting "32M" as the size of the benchmark. The
benchmark only uses a single core. The result should be
around 46 minutes runtime.

http://www.xtremesystems.com/pi/super_pi_mod-1.5.zip

The reason such a large test size must be used, is to defeat the
large cache on the i7. The Athlon has a smaller cache, and both
processors must be treated fairly. By running "32M", that means
the system memory is used to store intermediate calculation
results, making the benchmark more representative of real
world results (where the application software could be a
cache buster). A 32M test size, defeats the cache on both
processors.

HTH,
Paul

>
> The easy way to tell Passmark is comparing *all* cores, is to compare
> the performance of a Q9650 3.0GHz Quad core to an E8400 3.0Ghz Dual core.
> The quad is 4459. The dual is 2046. So the Passmark benchmark must
> be multithreaded, and comparing all available cores combined.

obviously

>
> SuperPI is single threaded. This is an Athlon XP 2600+ Thoroughbred @
> 2150mhz
> running SuperPI 32M. Execution time is 46min 59sec 940ms. (I just checked
> a table for processors, and 2600+ is listed as 2083MHz, so close enough.)
>
> http://hwbot.org/result.do?resultId=593780
>
> Intel Core i7 Processor Extreme Edition I7-97 at 3.33GHz. Well,
> it appears all the submissions for this processor, are overclocked.
> So this is running at 3.855GHz. Execution time is 9min 56sec 420ms.
>
> http://hwbot.org/result.do?resultId=850201
>
> We need to scale that result a bit. Convert to seconds first.
> 9min 56sec 420ms = 596.42 sec.
>
> 3855
> 596.42 * ---- = 690 seconds (if run at stock speed)
> 3330
>
> Now, convert the Athlon2600+ Thoroughbred time to seconds.
> 46min 59sec 940ms = 2819.94 seconds.
>
> Therefore, a Core i7 975 single core is 2819.94/690 = roughly 4.08 times
> faster.

that's sound reasonable, thanks for your calculus!

>
> You can validate the SuperPI benchmark on your own computer.
> Run this, selecting "32M" as the size of the benchmark. The
> benchmark only uses a single core. The result should be
> around 46 minutes runtime.
>
> http://www.xtremesystems.com/pi/super_pi_mod-1.5.zip
>
> The reason such a large test size must be used, is to defeat the
> large cache on the i7. The Athlon has a smaller cache, and both
> processors must be treated fairly. By running "32M", that means
> the system memory is used to store intermediate calculation
> results, making the benchmark more representative of real
> world results (where the application software could be a
> cache buster). A 32M test size, defeats the cache on both
> processors.
>
> HTH,
> Paul

humm i guess it's running on windows only, and i got no windows
installed.. i trust your links
Thanks!

"Pen" <(E-Mail Removed)> wrote in message
news:h92qp5$uhu$(E-Mail Removed)...
> blibloo wrote:
>> Ok just a simple question: how much faster will a SINGLE core of an core
>> i7 975 be vs an athlon XP 2600+ (2 Ghz) ? I suppose one get the core i7
>> with DDR3 and best mobo available for this cpu at this moment, and the
>> athlon is with DDR , NF7 mobo.
>> Do you guys have a precise idea? I can't find a benchmark with the athlon
>> XP 2600+ and the core i7 975...
>
> http://www.cpubenchmark.net/cpu_list.php
> Well you've selected the fastest cpu one the list to
> compare with an obsolete one. 15+ times faster.

But that is with 4 cores, so a single core would be one quarter of that
speed, or approximately 2x faster.