Athlon XP vs Sempron

S

Some Llama

Limp Gawd
Joined
Aug 6, 2001
Messages
379
so it looks like the only current difference is 256k of L2 cache, sempron has 256k and XP has 512k.

Can anyone help me with links that show benchmarks of the performance diff between the 2?

I am building a new system for a buddy (on a budget) and want to get a nforce2 solution but I am not sure if I should spend the extra cash for a XP over a sempron... thanks in advance!!
 
A Sempron is simply a badged retail Tbred XP with 256k L2 cache for the low end market. It's the same XP that was popular before the Barton. The Barton has 512k L2 cache and maybe more overclockability on it, but that's it.
 
Remember the old XP's with 266MHz FSB? That was the thoroughbred cores. They modified it and rebadged it as a value processor.
 
In my experience, 256k of L2 just doesn't cut it, at least in the world of socket A. Even at 400mhz fsb, i can see the lag in desktop stuff that isn't there with a barton core.

If you're going to get a socket A system, I say get a barton core. It is snappier all around. My real advice is to go for an A64 2800 though. Only a few bucks more, but worth it.

By the way, in case you didn't know, not all athlon XP's have 512k L2.
 
Get a mobile 2400+ and set it for 2.4 Ghtz. It's got 512 cache, overclocks like a beast, and runs cool(er).
 
In my experience, 256k of L2 just doesn't cut it, at least in the world of socket A. Even at 400mhz fsb, i can see the lag in desktop stuff that isn't there with a barton core.
Click to expand...


Bologna. I have tested many many Tbreds and Bartons and I notice no difference clock for clock in windows doing tasks, unless one is clocked higher.
 
Aren't there some Thorton core Semprons too though?

And aren't they bartons with half the cache disabled>?

For overclocking it seems the Thorton cores would prolly do almost as good as barton xp's only they are cheaper.
 
Since I'm getting a Sempron 2800+ for gaming (yes, I am poor), I did a lot of googling to try and find a review that compares the Sempron to the XP, so I know if I should consider using my Socket A to "upgrade" to one of the older XP chips. The only useful review with this comparison was on sharkyextreme. Apparently, the Sempron 2800+ performs about as well as a Athlon XP 2400, though depending on the application sometimes one or the other performs better.

The Sempron 3100+ should be compared to the A64s, since it's a Socket 754 and performs a lot better than the other Semprons.

Most reviews compared the Sempron to the CeleronD, blech. Of course kicked the Celeron's ass.
 
so what I was looking at was the 2800+ xp or sempron, I have a XP 2800+ myself and it is quite snappy, but if it really didn't matter I would go sempon for this new system to save 40-50 bucks (darn budget)...

I was also thinking a 2500+ mobile and OC to 3000 or ahap (as high as possible) since they OC so well.. are the 2400+ better for overclocking?

I don't want to go AMD 64 unless i can do it right (939 socket open ended upgrade path currently...) so 2800 AMD 64 wouldn't work since those are all socket 754.
 
foofighter06 said:
Aren't there some Thorton core Semprons too though?

And aren't they bartons with half the cache disabled>?

For overclocking it seems the Thorton cores would prolly do almost as good as barton xp's only they are cheaper.
Click to expand...
IMO, yes, and yes.

I don't really notice any advantages in ocing the Barton\Thorton cores over a good T-Bred B step. The only differences in process between the Barton and T-Bred B cores are the amount of l2 on the die - why would anything else matter when it came to overclocking?

In fact, it would seem more logical to bet the Barton overclocks wouldn't be as high because a lot MORE can go wrong with another few million transistors plus the electromigration at high clockspeeds. (course we know that's not the case)
 
but the XP series has always performed so much better wth more L2 cache comapred to it's counterparts (IE thorton) so wouldn't it be worth it to just get the higher L2 cache rather than even deal with OCing?
 
In benchmarks, the barton cores hardly perform better at all over the tbred/thorton cores.

Like I said earlier, I can notice a difference... Not in gaming or benchmarks (i'm not a gamer), but in desktop stuff, office apps, switching, multitasking stuff, the barton is able to do it better (smoothly) than the 256k L2 chips, which stutter a bit.

In gaming and benchmarks, I don't think there was ever a significant advantage to that extra cache.
 
GodsMadClown said:
Get a mobile 2400+ and set it for 2.4 Ghtz. It's got 512 cache, overclocks like a beast, and runs cool(er).
Click to expand...
I have to agreed with GMC a 2400 or 2600. A friend of mine had a 2600 and o'ced to 2.7ghz and I bought one just to do the same thing and for the price you can't beat it. Another friend of mine still has his 2400 that o'ced to 2.4ghz.
http://www.newegg.com/app/ViewProductDesc.asp?description=19-103-436&depa=1
http://www.newegg.com/app/ViewProductDesc.asp?description=19-103-439&depa=1
 
nylint said:
IMO, yes, and yes.

I don't really notice any advantages in ocing the Barton\Thorton cores over a good T-Bred B step. The only differences in process between the Barton and T-Bred B cores are the amount of l2 on the die - why would anything else matter when it came to overclocking?

In fact, it would seem more logical to bet the Barton overclocks wouldn't be as high because a lot MORE can go wrong with another few million transistors plus the electromigration at high clockspeeds. (course we know that's not the case)
Click to expand...

That's the theory on why the celeron, and the durons were such good overclockers. Less cache is less heat.
 
You must log in or register to reply here.
Back
Top