![]() |
|
|
#1 |
|
Jul 2004
Nowhere
809 Posts |
Im wondering is prime 95 optimized for the sempron proc.
|
|
|
|
|
|
#2 |
|
Sep 2002
2·331 Posts |
Two different semperon socket A and 754, both 32 bit, replace duron.
The socket A is like an Athlon XP with 256 KB L2. The 754 pin is like an Athlon 64 without the 64 bit modes with 256 KB L2. |
|
|
|
|
|
#3 |
|
Jul 2004
Nowhere
32916 Posts |
yes but if optimized shouldnt it show some increase.
|
|
|
|
|
|
#4 |
|
Oct 2004
232 Posts |
Note socket 754 doesn't support dual channel memory which is faster. Socket 939 does.
If you want to go with AMD I would steer clear of Semptron and go with an Athlon64. They have great benchmarks for trial factoring if you run them in the new 64 bit client. More flexibility more futureproofing and probably more resale value. Even the "3000+" model is not so much more expensive than Sempron. I would not purchase a 754 mobo/processor. |
|
|
|
|
|
#5 |
|
Jul 2004
Nowhere
809 Posts |
I bought a semptron its clocked at 1.5 ghz its a very nice proc. I know the semps roots are the A64 but im wondering if anything more can be squeezed from it.
|
|
|
|
|
|
#6 | |
|
Sep 2003
Borg HQ, Delta Quadrant
2·33·13 Posts |
Quote:
|
|
|
|
|
|
|
#7 |
|
"Jason Goatcher"
Mar 2005
3·7·167 Posts |
Just my two cents, but Semprons have a small cache. When I upgraded my computer, I wanted a faster processor specifically for doing more work on a DC project called Find-a-Drug. In my case, I feel the smaller cache size meant a cheaper processor, so it was a great deal for me. Also, Find-a-Drug's performance generally isn't affected by cache size.
The problem comes when you run a program like LLR, which LOVES large caches, then it may be worthwhile to get a processor with a bigger cache. It all depends on what you want to do with the hardware. In my case, I intend to build a Linux dvr and I'm almost positive it'll be some form of Intel chip I'll put in it for compression. But now I'm babbling.
|
|
|
|
|
|
#8 |
|
Jul 2004
Nowhere
14518 Posts |
intel pent 4 2.6 northwood with 800 mhz fsb for ur dvr. i use my semp for tf and running ltsp on linux :)
|
|
|
|
|
|
#9 |
|
Mar 2003
New Zealand
13·89 Posts |
A small cache is no problem for some computations. I have found that a 128KB cache Celeron is just as fast as a 512KB cache P4 for Prime95 ECM on numbers up to about 2^14000 or so. I expect something similar will apply to the Sempron.
2^14000 is too small for many prime search projects, but it is plenty for most factoring projects. A Celeron punches well above its weight in projects like LMH and many ECM projects, and I expect a Sempron would also do well at those where SSE2 is not currently so useful (LMH up to 2^64, ECM on numbers other than k*2^n+c for example). |
|
|
|
|
|
#10 |
|
Sep 2002
Austin, TX
3×11×17 Posts |
Like PrimeCruncher said, the Sempron is based off of 2 different cores.
Remember the AthlonXP? It has been renamed to Sempron. AMD did this, because they didn't want their "entry level" chips to carry the name of their more pricy chips, the modern Athlon64s. The original Pluto Athlon, Thunderbird Athlon, Duron, and AthlonXP are all K7 chips and all fit socket A (AMD's former universal plug). This means that any Sempron that fits socket A is automaticly a K7 and any optimization for the AthlonXP will also work for any K7 Semprons. Here is the kicker, Semprons are not exclusivly former K7 AthlonXPs. Some Athlon64s are cut down to the "entry level" catergory and are also called Semprons. These semprons are based on the K8 Athlon64 chips and are very different from the K7 Semprons. K8 Semprons do not fit on Socket A. Instead they fit on AMD's newer socket 754. |
|
|
|
|
|
#11 |
|
Jul 2004
Nowhere
809 Posts |
And i thought this was a dead thread
have socket a semprons changed any bit from aths
|
|
|
|