DFI Brings Dothan to Performance Desktop Computing

[perkam]

????

I actually cant understand the question but I can answer it I guess. According to the Processor Finder at Intel Website, none of the current line of dothans (1.5 - 2.0) will be H-PGBA -> They will ALL be PPGA.

However, coming first quarter 2005 probably, Intel will release the a dothan clocked at 2.10 Ghz which will be available both in PPGA and H-PBGA flavours.

Hope that helps abstrakt, and if it does, perhaps you can explain to the rest of us the difference between the two and their significance.

This is ALL I could gather up:

PPGA
Short for Plastic Pin Grid Array, PPGA was developed by Intel in 1993 as a solution to heat related issues with processors with more transistors than earlier processors using CPGA.

H-PBGA
Packaged Ball Grid Array - ???

Perkam

[Dojo | Warlord]

Yea, but what the hell is he talking about...

socket 855???

[perkam]

Aaah...I guess we're all a little confused in our own way...

855 (855GME to be exact) is the name of the Intel chipset that supports the Socket 479 Pentium M Dothans.

For example, 875 is the chipset that supports Socket 478 Prescotts and Northwoods and in the same way 855 was created for the 479 Pentium M.

Currently, only Aopen and DFI have desktop mobos for supporting the 855gme chipset. However, Intel will most likely go mainstream with the Dothans in mid-2005 - right now a Dothan system is not that feasible (dont tell Kyosen I said that though )

Perkam

[jjcom]

yeah I know what you mean. Dothans maybe fast, but if your like me just can't spend 600 bucks or so on a highend processor. But those things overclock like mad....
I just thought as to why Intel is waiting for Dothan to go mainstream. I just thought..."hey, ~22W @ 2.0ghz=very high qulity core" So I think that may explain why they cost so much.
They might be nice once they get on a 875 chipset. If only. I can dream can't I? Anyway...what are Intel's plans for these 533mhz FSB Dothans? Are they going to be on a 935 chipset or something? I know their on a different one with...what is it? PCI-E, DDR2...both nice additions to the already powerful core. Somehow I think the chipset is limiting that processor...I find that a scary thought. Just think what one of those would do on a 875 board. We'd be seeing 1066mhz FSBes....that would be very nice to see

[abstrakt]

Yea, but what the hell is he talking about...

socket 855???

lol, just read the Intel processor spec finder wrong.

[charlie]

currently Dothan 2.0 is only $400

[cpulloverclock]

Aopen is back
http://download.aopen.com.tw/userdow...l=i855GMEm-LFS

a new bios


File Download Property
Products Motherboards(Intel Pentium M: Socket 479)
Models i855GMEm-LFS
Categories BIOS Version R1.01a Beta
File Name wi85101a.zip File Size 512.9KB
Software Name N/A
Descriptions
Improved system performance.
please use formal release if occured any instable condition

OS Support WinXP, Win98SE, WinME, Win2000, DOS
Released Date 11/3/2004

[Siphilon]

a new bios


File Download Property
Products Motherboards(Intel Pentium M: Socket 479)
Models i855GMEm-LFS
Categories BIOS Version R1.01a Beta
File Name wi85101a.zip File Size 512.9KB
Software Name N/A
Descriptions
Improved system performance.
please use formal release if occured any instable condition

OS Support WinXP, Win98SE, WinME, Win2000, DOS
Released Date 11/3/2004

Sweeeet....the AOpen board is now avail. in the USA

Looks like I will be giving it a try until I can buy the DFI board to compare.

currently Dothan 2.0 is only $400

WHere? Retail box?

[Shade00]

Sweeeet....the AOpen board is now avail. in the USA

Looks like I will be giving it a try until I can buy the DFI board to compare.

Where is it available?

WHere? Retail box?

The Dothan 1.6 is around $200, and the Dothan 2.0 is about $440 retail at Newegg.

[kyosen]

> right now a Dothan system is not that feasible (dont tell Kyosen I said that though )
--
Hi, I'm here

> Descriptions
> Improved system performance.
--
I've tried it, but there is no significant difference, I feel...
...at least, at SuperPI
The only notable thing which I've noticed so far is that
multiplier setting in the BIOS became effective.


Yesterday, I touched DFI's sample board for a while.
My first impression is, the performance seems almost
same as AOpen's board...at least, at SuperPI again
Memory interleave is not available on DFI' board, it seems.
So I feel that 25s@2.8G is difficult and mysterious...

[perkam]

The only notable thing which I've noticed so far is that multiplier setting in the BIOS became effective.

You mean on the DFI and not on the Aopen ???

[cpulloverclock]

> right now a Dothan system is not that feasible (dont tell Kyosen I said that though )
--
Hi, I'm here

> Descriptions
> Improved system performance.
--
I've tried it, but there is no significant difference, I feel...
...at least, at SuperPI
The only notable thing which I've noticed so far is that
multiplier setting in the BIOS became effective.


Yesterday, I touched DFI's sample board for a while.
My first impression is, the performance seems almost
same as AOpen's board...at least, at SuperPI again
Memory interleave is not available on DFI' board, it seems.
So I feel that 25s@2.8G is difficult and mysterious...

There will be some tweaks in the future I think to improve the perf, so to improve the bandwidth
dont worry about

[cpulloverclock]

Kyosen the throttling is off in the aopen board?

28s at 2.66GHz
33s at 2.6GHz
25s at 2.8GHz

weird

[cpulloverclock]

there is a difference of 3-5s between the japanese and english version with the dothan???

[kyosen]

>> The only notable thing which I've noticed so far is that
>> multiplier setting in the BIOS became effective.
--
> You mean on the DFI and not on the Aopen ???
--
I mean on the AOpen.
...Confusing is thanks to cpulloverclock?

> there is a difference of 3-5s between the japanese and english version with the dothan???
--
No, difference is 1 second at most

[perkam]

Hey kyosen, where are your benchies for the Dothans on the DFI and Aopen.

I wouldnt mind some 3dmark benchies at 2.5 or 3.0Ghz Dothan

[kyosen]

Hey kyosen, where are your benchies for the Dothans on the DFI and Aopen.

I wouldnt mind some 3dmark benchies at 2.5 or 3.0Ghz Dothan

I have no bench rusults on DFI board.
But don't you know these results on AOpen board?
http://219.113.251.125/c-board/c-boa...ne;no=3075;id=

Unfortunately, XT-PE's 3DMark2001 score is not high
Game2 score is very low...I suspect that this is Dothan's peculiarity.
http://219.113.251.125/c-board/c-boa...ne;no=3085;id=

[kiko]

kyosen does OCZ booster go up to 3.2v w/o vdimm mode on aopen?
and which font do you and your japanese friends use for crystalcpuid?

Ciao

[kyosen]

> kyosen does OCZ booster go up to 3.2v w/o vdimm mode on aopen?
--
I confirmed 3.5V over by only OCZ booster, without Vddr-mod, on AOpen.

BTW, be careful...I strongly recommend NOT to use OCZ DDR booster
on DFI's i855 board(& also 875P-T)...you'll see burning booster

> and which font do you and your japanese friends use for crystalcpuid?
--
In my case, quite oridinary "MS UI Gothic(9pt)"...any problem?

[ISS]

I think Dothan for desktops is the greatest threat, the Pentium-M Centrino 2mb cache. some manufacturers having already announced shipping them!

DFI in particular.

here is some Pentium-M benchmarks. Look below in Figure 1.

DFI news
Figure 1

I find that the Doom3 results are rather strange. At 640 x 480 LQ, it should be CPU bound and not gfx card. However, I am rather sruprise that A64 4000 only manages a paltry 87.8 FPS. The test did not state what gfx card they are tested with though. Neither did they state which demo they are using.

Anyway, just to add that its impossible for Dothan to out perform A64 in apps where FPU is the main factor. The FPU is derived from P3 (which itself is a revised P Pro) and Intel did not make any significant enhancements to it (since its a mobile CPU), thus in apps where FPU power is important and memory bandwidth dependent (rather than L2 cache), Dothan will not be able to out perform A64.

Super PI is known to be a memory + L2 cache depedent program rather than FPU dependent program (which explains why P4Es run so well, also why Dothan with 2MB L2 rules).

[kiko]

> kyosen does OCZ booster go up to 3.2v w/o vdimm mode on aopen?
--
I confirmed 3.5V over by only OCZ booster, without Vddr-mod, on AOpen.

BTW, be careful...I strongly recommend NOT to use OCZ DDR booster
on DFI's i855 board(& also 875P-T)...you'll see burning booster

> and which font do you and your japanese friends use for crystalcpuid?
--
In my case, quite oridinary "MS UI Gothic(9pt)"...any problem?

:/ i do not like burning other than a CD
10x for the tip.
For crystalcpuid if you are using that font...i have some problem.
W/ that font my crystal shot is bigger than superpi...
Your japanese shots are so small. Only font near that size is lucida console but numbers are not in a good position.

Ciao

[cpulloverclock]

ftp://ftp.x86-secret.com/dfi/855gme/BIOS/X86F03.zip

** Changelog **

- Remove CPUTest86+ (Not ready yet)
- Fix PCI Freq displayed incorrectly
- Changed Memtest86+ to v1.30
- Fix a bug with some PC4000 & PC4400
- Added some new POST Codes


http://forum.x86-secret.com/viewtopic.php?t=2614

[misteroadster]

I think Dothan for desktops is the greatest threat, the Pentium-M Centrino 2mb cache. some manufacturers having already announced shipping them!

DFI in particular.

here is some Pentium-M benchmarks. Look below in Figure 1.

DFI news
Figure 1

This graphic is wrong , i'm sure , with my FX53 Cat 4.10 ,Sgle channel 15X170 (2550mhz) Agp 4X , X800XTPE 520/560, i'd done 94 Fps at 2nd run , ive deleted the Doomcfg before the test.

[Ailleur]

This graphic is wrong , i'm sure , with my FX53 Cat 4.10 ,Sgle channel 15X170 (2550mhz) Agp 4X , X800XTPE 520/560, i'd done 94 Fps at 2nd run , ive deleted the Doomcfg before the test.

Right, surely we can draw conclusions from the results of two different pcs... not
And Samuel stated those were "first run" results.

[uclajd]

A new review by Game PC.