SocketA Sempron 2600 faster than S754 Sempron 3100+

[Brian K]

They both run the same clock speed, both running DDR400 memory.

Every CPU benchmark I've run so far (and glxgears) shows the 2600 leading
the 3100 by a small amount.

Any ideas what's up?

Brian K

 

[Wes Newell]

They both run the same clock speed, both running DDR400 memory.

Every CPU benchmark I've run so far (and glxgears) shows the 2600 leading
the 3100 by a small amount.

Any ideas what's up?

glxgears doesn't care aobut the cpu much at all. It basically test 3D on
the video card. And there's a wide range of speed of it with even GF2
cards beating FX cards running glxgears. So, what's up is you aren't
running a cpu benchmark.

Go here and select the second Sempron 3000+ (socket A barton core) on the
list as 1 cpu, and then as the 2nd cpu select the Sempron 3100+ (754) and
compare them. Selecting them just highlights them in red. Each benchmark
shows all CPU's. Now the socket A sempron 2600+ isn't on the list, but I
think you'll conceed The socket A Sempron 3000+ is faster and more than a
fair comparison used as a replacement for the 2600+.

http://www23.tomshardware.com/cpu.html

 

[* * Chas]

glxgears doesn't care aobut the cpu much at all. It basically test 3D on
the video card. And there's a wide range of speed of it with even GF2
cards beating FX cards running glxgears. So, what's up is you aren't
running a cpu benchmark.

Go here and select the second Sempron 3000+ (socket A barton core) on the
list as 1 cpu, and then as the 2nd cpu select the Sempron 3100+ (754) and
compare them. Selecting them just highlights them in red. Each benchmark
shows all CPU's. Now the socket A sempron 2600+ isn't on the list, but I
think you'll conceed The socket A Sempron 3000+ is faster and more than a
fair comparison used as a replacement for the 2600+.

http://www23.tomshardware.com/cpu.html

Wow! That's an impressive table.

Chas.

 

[Brian K]

Wow! That's an impressive table.

Chas.

And I've tried a number of CPU benchmarks that show the Sempron 2600 still
edging over the Sempron 3100...

Brian K

 

[Wes Newell]

And I've tried a number of CPU benchmarks that show the Sempron 2600 still
edging over the Sempron 3100...

Then your benchmark results are flawed by other factors of the different
platforms or misconfiguration. The 754 Sempron 3100+ runs stock at 1.8GHz
with 256K L2 cache (roughly a remarked A64 2700+). The Socket A Sempron
2600+ runs stock at 1.83GHz with 256K L2 cache also (roughly a remarked
2200+ XP). The 754 Sempron runs instructions at about 1.75 to 1.45 over
the socket A Sempron. The only benchmarks that that it would win would be
ones to determine raw clockspeed. If I'm not mistaken the socket A Sempron
3000+ at 2GHz with 512K cache doesn't even beat the Sempron 3100+ in a
single benchmark of the 20 or so listed on tomshardware. There is one
listed in Farcry, but that's a mistake, unless yoyu think the 3000+
Sempron is way fatser than an Athlon XP 3200+. The Sempron 3000+ does win
a few of the meaningless benchmarks that only measure core speeds since
it's clocked 200MHz higher, but it's dead in the water on everything else.
The socket A Sempron 2600+ wouldn't win a single benchmark and would be
severely beaten by the 3100+ in all but raw speed test.

 

[Brian K]

Then your benchmark results are flawed by other factors of the different
platforms or misconfiguration. The 754 Sempron 3100+ runs stock at 1.8GHz
with 256K L2 cache (roughly a remarked A64 2700+). The Socket A Sempron
2600+ runs stock at 1.83GHz with 256K L2 cache also (roughly a remarked
2200+ XP). The 754 Sempron runs instructions at about 1.75 to 1.45 over
the socket A Sempron. The only benchmarks that that it would win would be
ones to determine raw clockspeed. If I'm not mistaken the socket A Sempron
3000+ at 2GHz with 512K cache doesn't even beat the Sempron 3100+ in a
single benchmark of the 20 or so listed on tomshardware. There is one
listed in Farcry, but that's a mistake, unless yoyu think the 3000+
Sempron is way fatser than an Athlon XP 3200+. The Sempron 3000+ does win
a few of the meaningless benchmarks that only measure core speeds since
it's clocked 200MHz higher, but it's dead in the water on everything else.
The socket A Sempron 2600+ wouldn't win a single benchmark and would be
severely beaten by the 3100+ in all but raw speed test.

Theoretically, yes, my Sempron 3100+ SHOULD be outperforming my Sempron
2600+, I agree completely. Unfortunately, it's just not happening on my
hardware for whatever reason...

Byte Unix Bench dhry2 shows the 2600 at an index of 174.5 vs. the 3100 at
171.5, mencoder shows the 2600 runs about 5fps faster at encoding vhs
captures to DVD, and the 2600 beat the 3100 in a kernel compile also...

Some of that might be due to a faster hard disk in the 2600, maybe, but it
still leaves me scratching my head. I've even got the faster RAM in the
3100! Quite possibly is the cheapass motherboard I got with the bundle
(from Fry's), an ECS Nforce3-A, but I really am at a loss.

Not that I'm really complaining, it's still faster than I needed to begin
with, but it's just odd...

Brian K

Brian K