Official Gigabyte X48T-DQ6 Info Thread

[Supershanks]

yes, i had similar problems with this board at and above 520 FSB with all of the bioses running Super PI 32M...it must be something to do with us being at the very edge of memory specs and trying to find the right memory settings to make the chipset and cpu compatible with running super pi...all my video benches were completed but i couldn't run 32M bench get that one

I think at the minute it's the CPU as I showed in the link , can do it with the E8400 not with the E8500 which is a little odd as the E8500 is binned higher. Maybe my 8400 is a good one that got away

[Ace-a-Rue]

Originally Posted by Supershanks

I think at the minute it's the CPU as I showed in the link , can do it with the E8400 not with the E8500 which is a little odd as the E8500 is binned higher. Maybe my 8400 is a good one that got away

binned higher with a point '5' higher multi.

some are luckier (Luck) than others when you get the GOOD chip and others get an 'ok' chip...hehehehehe

[X4er0]

Supershank and Ace what kind of memeory are you guys using?

[Supershanks]

some are luckier (Luck) than others when you get the GOOD chip and others get an 'ok' chip

Yep think my E8400 is looking good whilst the E8500 is ok.


CellShock PC3-14400 (DDR3-1800) 2GB DDR3 Great stuff

[Ace-a-Rue]

OK...going to put in my 9800GTX card to see what PCIe 2.0 specs do for me with the X48 board...the 8800GTS has 320 mb of video memory and the 9800 has 512 mb..additionally, it is a faster card...i will have comparison slides to compare to the 8800GTS...all system memory settings will be identical between the two cards at 500 FSB.

[Ace-a-Rue]

here are the simplified spec differences between my two video cards:

8800GTS:

GPU GeForce 8800GTS
Core clock 580MHz
Stream Processors 96
Memory
Memory Clock 1800MHz
Memory Size 320MB
Memory Interface 320-bit
Memory Type GDDR3
RAMDAC 400 MHz
Max Resolution 2560 x 1600

9800GTX:

Interface
Interface PCI Express 2.0 x16

GPU GeForce 9800 GTX(G92)
Core clock 675MHz
Stream Processors 128
Memory
Memory Clock 2200MHz (effective)
Memory Size 512MB
Memory Interface 256-bit
Memory Type GDDR3
RAMDAC 400 MHz
Max Resolution 2560 x 1600

[Ace-a-Rue]

here are the 4 video benches using the 8800GTS:

[Ace-a-Rue]

i realize this is not the best comparison since the card has different specs as far as core clock speed, stream processors, memory clock speed and memory on the cards:

EDIT:...the card is not overclocked and the driver was set to high performance.

[Ace-a-Rue]

not as good as my IP35 PRO board...two reasons why...1st, E8500 versus QX9650 on the PRO board...2nd, single drive operation versus Raid 0.

going to put in my QX9650 and set up Raid 0 with Vista 64...the difference between XP SP2 and Vista 64 on the Pro board was 500 points more for Vista64.

be back in a few hours...

[Ace-a-Rue]

ran out of time last night...i yanked the heatpipe system to check how the TIM (grey stuff) was interfacing PWM, NB and SB...to my surprise, it wasn't even melted or tacky!...i could peel off the thin layer of TIM without any problems at all...i polished the NB and PWM heatsink bottoms and then applied AS Ceramique...now, i feel some warmth coming into the heatpipe system at the PWM heatsink.

i have a 70mm fan on the NB; so far that is it.

working on installing SP1 for Vista64...i will be getting to PCMark2005 very soon...i am curious now since PCMark05 would not run without SP1...it ran on my IP35 Pro board with SP1 installed....the 1.20 patch is suppose to make it compatible with vista; obviously not without SP1.

[Ace-a-Rue]

the only difference is the Intel chipset in Raid 0 is not as good as the Areca Raid card which has its own memory and basic cpu processor.

even though the the ggia board scored about 1800 points higher the video bench the overal PCMark05 score is only about 325 points higher....SO....to get a true comparison between chipsets, i will remove the Areca Raid card from this puter and move it over to the giga board and run setup again with vista 64 (SP1 after the OS is installed).

notice the memory sub score and video sub score are better than the Pro board sub scores.

1st slide show the results from the IP35 Pro:
Vista64 w/SP1
QX9650
9800GTX
2 x 2gb DDR2 system memory
Areca Raid 0 (128K Stripe) with 2 x 250gb Seagate one platter drives

2nd slide show the results from the GA-X48T-DQ6:
Vista64 wSP1
QX9650
9800GTX
2 x 2gb DDR3 System memory
Intel Raid 0 (128K Stripe) with 2 x 320gb Seagate 2 platter drives

[Supershanks]

Great work Ace, very interesting. Good to see a pukka comparison, thanks

[Ace-a-Rue]

Originally Posted by Supershanks

Great work Ace, very interesting. Good to see a pukka comparison, thanks

the IP35 Pro is a VERY good board...PCMark2005 is a good benching prog that shows the overall performance...i think hard drive performance and CPU performance have a strong scoring weight factor, and video and memory follow those two.

here is the GA X48T-DQ6 with the Areca 1210 model Raid card using the same drives that i used on the IP35 Pro...i had four of these.

Vista64 was installed and then, it was updated with SP1.

i had the 9800GTX running at default speed with only the driver set to high performance...the bios had the video accelerator set to auto.

the FSB was set eaxctly the same as it was on the Pro board and the same specs for cpu, QX9650...the only difference was the chipset and memory; being DDR2 versus DDR3...ddr3 is obviously has a higher Mhz threshold but also is accompanied by higher latencies and timings.

we see the Giga X48T board move out in front of the P35 board by a good margin..a thousand points!

[Supershanks]

At lest we've not been throwing our money around to stand still
good find. That's a pretty good performance from the Areca Raid card Hadn't considered them too much before. may reconsider

Interesting comparison:-

Memory - we know DDR2 to DDR3
Graphics - pcie 2 I guess , hadn't expected that bigger lift but ok
HDD - Not quite got my head round this one ??

[dinos22]

where do you guys get these bioses
i can email my GB contact but i'll probably get an answer tomorrow and i want one NOW hahah

pls point me in the right direction

also would anyone have this bios
have a look at the efficiency EEK

F3F bios
http://www.hwbot.org/result.do?resultId=685943

[Ace-a-Rue]

compliments of BenchZowner: http://www.benchzone.com/Downloads/B...e/GA-X48T-DQ6/

[Ace-a-Rue]

Originally Posted by Supershanks

At lest we've not been throwing our money around to stand still
good find. That's a pretty good performance from the Areca Raid card Hadn't considered them too much before. may reconsider

Interesting comparison:-

Memory - we know DDR2 to DDR3
Graphics - pcie 2 I guess , hadn't expected that bigger lift but ok
HDD - Not quite got my head round this one ??

WOW!...EXCELLENT chart...you are hired!

that makes it easier to see the increases coming from the X48T

EDIT:...if you think about the hard drive score...the PCMark2005 suite was written for one hard drive...now, most people do not expect double your return going from single to Raid 0...usually you expect, maybe, a 30% increase not double.

Intel's solution uses the system's CPU and memory...so...it has to wait in queue for its turn to process data....the Areca card that i have has its own processor (330 Mhz) and 256 MB of memory...no waiting on system memory or CPU for computation and storage of data like you would see with Intel's solution.

[Supershanks]

WOW!...EXCELLENT chart...you are hired!

just a screenshot of something knocked up in a couple of minutes in Excel

Yep so the HD score is a direct function of faster memory performance say 5% , yep can buy that

So if we eliminate the DDR3 impact :-
is 404+636 , are we saying gain is all down to memory, it'a a pity the total isn't additive. It's not an average either (Sum element score & / by 4) . We could do with someone with a X48 DQ6 to repeat your tests

Anyway Gigabyte rules OK

[Ace-a-Rue]

Originally Posted by Supershanks

just a screenshot of something knocked up in a couple of minutes in Excel

pat yourself on the back...you did an excellent job with layout and color scheme!...simple, YES, but it is a good analytical tool.

Yep so the HD score is a direct function of faster memory performance say 5% , yep can buy that

well mate, i am not sure how to take your statement...if you strictly mean taking a single drive from a P35 board and moving it over to an X48 board, yes, a 5% increase in performance due to DDR3 memory performance.

the reason Areca does better is two fold: it has it own processor to deal with requests and its own onboard memory to store data, therefore, alleviating any bottleneck with system memory and system cpu....therefore...a nice boost in speed for those acquiring enthusiasts!

So if we eliminate the DDR3 impact :-
is 404+636 , are we saying gain is all down to memory, it'a a pity the total isn't additive. It's not an average either (Sum element score & / by 4) . We could do with someone with a X48 DQ6 to repeat your tests

NOPE...a good gain is going from the P35 to X48 chipset...i think the gain from X38 to X48 is going to be around 2.5%.

don't forget to factor in the PCIe 2.0 spec and it is even better going from P35 to X48.

Anyway Gigabyte rules OK

IT DOES!

[Venom-Crusher]

My result with E6300 & E8500 (normal water cooling)

http://www.vmodtech.com/modules.php?...owpage&pid=334

[Supershanks]

well mate, i am not sure how to take your statement...if you strictly mean taking a single drive from a P35 board and moving it over to an X48 board, yes, a 5% increase in performance due to DDR3 memory performance.

Sorry didn't make myself clear , the Areca configuration & performance is identical on both the P35 & X48. The Card's onboard CPU & memory are common in both cases. Yet there is a nearly 6% 636 point improvement in X48. I agree with your supposition x4 restriction on the P35 is a reason for the improvement,

Great clocks Venom-Crusher
On the E8500 your temps look particularly good as your showing the wrong Tj Max see Unclewebb's Realtemp which meands youer temps are probably 10C less than shown
ie around 66C

luck

[Ace-a-Rue]

Originally Posted by Venom-Crusher

My result with E6300 & E8500 (normal water cooling)

http://www.vmodtech.com/modules.php?...owpage&pid=334

that is some pretty amazing memory coming from OCZ....those speeds and timings are just awesome for DDR3!

[Venom-Crusher]

Originally Posted by Ace-a-Rue

that is some pretty amazing memory coming from OCZ....those speeds and timings are just awesome for DDR3!

Thanks, but my memory is cellshock ddr1600. I used spd of OCZ DDR1600 XMP memory for it.

[Ace-a-Rue]

aaaah!...no wonder why such GREAT numbers!

i immediately started looking for that (OCZ) memory...glad you told me, otherwise i would have bought OCZ...Supershanks has cellshock also but he had a bit more of a problem getting the timings in your range.

EDIT:...what did you use to flash cellshock with OCZ template?

[Venom-Crusher]

I used SPDTool 0.62. It not supported ddr3 to edit, but it support to flash.

I post it for you, Try it with SPDTool