Official AMD Sempron review out !

[Sheik]

Anandtech has a review of the 3100+ 754 and 2800+ Socket A

http://www.anandtech.com/cpuchipsets...spx?i=2139&p=1

It rules the Celeron D of course

Definitely going into my spare rig.

[JNav89GT]

good read

I predict alot of boxes housing the 3100+ for budget gamers

[reject]

i will never go back to 256k l2.
'AIUHB' 'Y' do i hate you?

[Shade00]

Bah, methinks AMD is going to sell a lot of the 754 Semprons.

[LilGator]

Not bad, but now with 2800+ A64s going for $140 or so, the high-end Semprons (3100+) might not be a good idea. It would be cool if they had S939 semprons, then you could throw a $39 2200+ chip in a board whilst you save up for an FX53

I'd like to see a head-to-head comparison with benches of a Sempron 2800+ versus a Mobile Barton 2600+ both pushed to their max clocks. Both similar prices. I wonder if for the same price, the Sempron replaces the mobile bartons, and nForce2

[Sheik]

Hardwarezone has done a little overclocking as well.

http://www.hardwarezone.com/articles...03&cid=2&pg=17

What fun is there if you don't overclock a processor to see where its limits lie? In this particular test, we increased the system bus frequency to about 250MHz. Together with its multiplier at 9x, it yielded a final core frequency of about 2.25GHz. Compared with the original clock speed at 1.8GHz, we managed to push the processor's speed by as much as 25%. However, in order to do that and yet maintain a stable system, we had to change the hard drive to an IDE-based component since Serial ATA drives are extremely sensitive to high bus speeds. In addition, we also lowered the Hyper-Transport frequency to a lower ratio and it resulted in an odd but stable speed of 750MHz. Finally, the memory divider was also lowered and yielded a final clock speed of about DDR433.

You can see that with a carefully planned overclocking scheme, one can typically achieve very good processor speeds with the AMD Sempron 3100+. The results here are quite encouraging as we managed to put the Sempron 3100+'s performance on-par with the Athlon 64 3200+ processor.

[althes]

not bad

[JWB]

So do these have locked multis? I guess no one knows yet seeing as they are not out. It would be cool to see them have their multis unlocked downwards like the clawhammer/newcastles.

[Sheik]

Before we tested on the EPoX 8KDA3+, we were sure to update the board's BIOS. When we accessed the "Power BIOS Features" section for overclocking, we found that the CPU Ratio Control was available. We were pleased to find that it was, in fact, adjustable and working, but the choices were only 8x and the default 9x. While this isn't a huge selection, we'll take it. The 8x setting dropped the CPU speed to 1600MHz with the FSB register set to 200, leaving a lot of room for increasing the FSB without overtaxing the processor.

http://hothardware.com/viewarticle.cfm?articleid=559

Better than nothing

[Shade00]

If they have Cool n' Quiet, they'll have unlocked lower multis.

[JWB]

Originally posted by LilGator
It would be cool if they had S939 semprons, then you could throw a $39 2200+ chip in a board whilst you save up for an FX53

Looks like there will be

http://www.theinquirer.net/?article=17499

To quote EPOX:

The firm said that the Sempron will support multiple sockets including A, 754 and 939.

[Shade00]

Yeah, they will have 939 Semprons, but not for a while, and they won't be $39. The 2200+-2800+ Semprons are only for Socket A.

[JWB]

How will we know which Sempron is for which sockter apert from them advertised on the retailers site? Do you know if there will be any code etc to denote this?

[Shade00]

Well... I don't really know. I do know that Socket A Semprons stop at 2800+, but that's about it.

[NivLac]

Originally posted by LilGator
I'd like to see a head-to-head comparison with benches of a Sempron 2800+ versus a Mobile Barton 2600+ both pushed to their max clocks. Both similar prices. I wonder if for the same price, the Sempron replaces the mobile bartons, and nForce2

>) yea man I'd lub to seee those too. i wonder how they'll preform when pushed to the max... now like is the ratings actual preformance when compared to a barton or is it just a number?!

like is a 2800+ sempron comparable to a 2800+ barton?? (sry im jw here...)

[Jrocket]

The performances ratings for the semprons are supposed to be compared to that of a celeron D I believe.

[Sheik]

Some more Sempron details.

LONDON — The Sempron processor line launched by Advanced Micro Devices Inc. (AMD) on Wednesday (July 29) has much in common with the Athlon XP line of processors, and in some cases may be the same die previously shipped as Athlon XP.

"There is overlap between the two product lines," said Bahr Mahony, a mobile marketing manager with AMD, who conceded that in some cases the Sempron processors could be the same die as Athlon XPs, with different clock frequencies specified.

The Athlon XP family was originally launched in October 2001.




The Sempron series uses at least three different die, with some based on the Thoroughbred processing core, others on the Barton core while the high-end desktop Sempron is based on the Dublin core. The desktop version also uses the hypertransport bus to memory rather than a front-side-bus, Mahony said.

Sempron processors are made using a 0.13-micron manufacturing processes. So-called "Mobile Sempron" processors are based on the Dublin core, which only surfaced in one Athlon XP processor, Mahony said.

Asked the difference was between Sempron and previously announced Athlon XP processors, Mahony cited a combination of features such as on-chip cache size along with the pairing of the guaranteed frequency of operation for the main processor and the the front-side bus.

"There we certain FSB frequencies we did not end up shipping in Athlon XP," said Mahony. Athlon XP processors have FSB frequencies of 266 and 333 MHz.

Mahony said different cache sizes could be achieved by changing the way on-chip cache was connected. Or, he added, a chip could be laid out again and a new die created to minimize die area and cost.

Source.
http://www.eet.com/semi/news/showArt...cleID=26100623