Biostar T-Force 6100-939 mobo?

[kamakazibond]

Hello..

I was wondering if anyone has used any of the Gskill memories on this motherboard.. T-Force 6100-939 (nf4)..

It is basically Biostar's attempt to break into the higher end market.. It is quite similar to many DFI mobo's including Memtest in BIOS, a BIOS Genie type of option to save up to 50 different settings, a +3v jumper to enable +3v to the memory, UV reactive pci/pcie slots, etc...

Basically I was wondering if anyone has tested some of the Gskill kits with this mobo to see what kind of performance I will be able to get out of it.

If the Gskill Rep would like to try and get ahold of one of these mobo's to do testing on, that'd be sweet!!

I can say that I have only been able to get my 3200PHU2-2GBZX to around 236MHz @ 2.5-3-2-5 2.6v or so with only minimal DOS based memtest testing.

Quick question for the Gskill tech, and I'm sure he prolly gets asked this alot..

Why do you recommend using windows memtest? When even the creator of the app tells us that while in windows you are unable to test all of the memory, and that he offers a DOS based version as well if you want to be able to test the entire amount of memory in the system.


In reference to this post, http://www.xtremesystems.org/forums/showthread.php?t=89366, Why do you recommend using Prime95's Large FFTs test, when it only uses 8MB of memory?


Well, ok thankx for any help..

kamakazibond

[GSKILL TECH]

i didn't try this mobo yet but i will if i get the sample one :)
the dos memtest still have little problem with CPU motherboard and ram in my point of view. i have many cases that dos memtest86+ pass 4~5hr complete test but can not even boot into windows..... why? becasue the timing is too tight but howcome memtest86+ didn't give me error? or for BH-5UTT say, DDR540 3.6v 24/7 machine, never bosh on me but it crush memtest86+ when it stated. need to power off the all system then reboot.
to me i believe windows stable is better than memtest86+ stable. but that cause a lot more time.!

[kamakazibond]

Thankx soo much for the response.. I do hope you can get ahold of this mobo, cuz it's actually quite nice.

I got some 2GBHZ today so I'll be doing some testing for a bit here.. we'll see what happens.

As far as the memtest86+ thing.. I would have to agree that I have seen some strange occurences, but that dosen't make me want to scrap the program entirely..

I would also have to agree with windows stable is better than memtest86+ stable, but I would re-word the sentance to say prime95/superPI stable inplace of windows stable :D

So I hope to pull 300MHz outta these 2GBHZ's we'll see :D wish me luck!!

kamakazibond

[kamakazibond]

Well so far I'm kind of disappointed..

I received my 2GBHZ kit yesterday. I did some prelimary testing using the built-in BIOS memtest (dos), which would error out at around 260-265MHz, primarily using test 5 looped, which from what I have been told is the most stressfull.

So I figured I would try out the recommended Windows OC tests. I opened up prime95 and started the large FFT test, then opened windows memtest which told me to run it a second time and set it to 825MB per test. I then opened CPU-Z, the Biostar OC utility and the Biostar Hardware monitor.

In BIOS I have set the CPU and other settings to default including setting the memory set to 200MHz (1:1), CPU multiplier to 9x (default for my 3000+ venice), and FSB to 200. Effectively undercklocking the memory to 200MHz.

I started to up the FSB while the tests were running, waiting 20 seconds or so between upping the FSB in one step increments. I had made it up to 260MHz, and I let it run like that for a while.. Well before too long one of the windows memtests popped up with an error, so I bumped the FSB down one notch and restarted that memtest. Again after not too long the other memtest popped up with an error.. I continued this untill I was at 256MHz FSB, which ran for quite a while without any errors..

So according to your windows OC instructions, I should lower the FSB down 5MHz, and run the full tests.. Which would put me at 251MHz... Hardly an OC at all..

All tests were done at 3-4-4-8 timings, also matching your posted timings for DFI motherboards, as much as possible, A64Tweaker in windows basically has all of the memory settings that the DFI mobo's have in BIOS, which again were matched to the timings you have posted.

Is there anything I am missing?? This mobo has a NF4 chipset on it, could there possibly be THAT much of a difference with the DFI mobo's??

Originally I had purchased a kit of 2GBHZ's for a MSI NF3 mobo, due to UPS messing up my order I ended up with 2 kits of 2GBHZ, one kit had a stick that would not POST at all, while the other kit would POST, but wasn't stable even at 200MHz. So I ended up buying some 2GBZX's for that MSI NF3 mobo, which was stable at stock 200MHz 2-3-2-5, but wasn't stable past 230MHz (on the MSI NF3). I then purchased this Biostar mobo, with NF4 on it.. The 2GBZX's seemed to do a tiny bit better on it reaching around 236MHz stable. So I ordered some more of the 2GBHZ's to see if they would run at least at 250MHz on this new mobo. Which they appear to be doing, but with everyone I see all over the place online saying that 270MHz should not be a problem, I am kind of disappointed.

I have a friend who just bought a DFI Ultra-D mobo and some 2GBHZ's comming over tomorrow to do some more testing.. I guess I'll find out if these sticks are just not good OC'ers or if it is the mobo that wont let me OC it..

BTW both times I bought the 2GBHZ's from newegg.com, they had an instant rebate for around $30.00. Is newegg lowering their price because these sticks aren't as good as previous ones?? Am I being unreasonable to expect to OC at least somewhat close to what 90% of others I see online are??

Any Ideas?? I surely welcome them!!

Oh BTW CPU-Z shows the timings by SPD are 3-5-5-10. Any reason for that?? Should I try those timings to see if I get any better results???

OK I just read the post where GS Tech had said that some people had issues with the SPD set to 3-4-4-8 being picked up as 3-3-3-6 so they have put the SPD to 3-5-5-10 so mobo's wont have any issues POSTing at 250MHz. I think I will still try these timings to see if it helps, at least from 8 to 10. The Xbit review of this memory I saw linked in another thread here said they were testing at 3-4-4-10.

Thankx for any help!
kamakazibond

ps- Mr. G.Skill tech, I have added u to my MSN but I haven't seen you online to chat with as of yet, I live on the East coast and usually am up late, if you get the chance to chat feel free.. My MSN is jasondev@bellsouth.net

Thankx again

[kamakazibond]

updated my signature

[kamakazibond]

OK so this next post here is GIANT.. I am gonna split it into 3 posts not including this one..

Sorry for such a long post, but I try to be as descriptive as I can so that there is no confusion about what has been done and what results I have seen..

So here it goes, the DFI Ultra-D / T-Force 6100-939 Testing Saga...

[kamakazibond]

OK here goes.. I'll try to remember most all of what we did..

I'll start 1st by saying that a friend came over on sunday with his brand new DFI Ultra-D mobo, and his Brand new (All from newegg BTW) GBHZ Dual channel mem kit.

Up until this he had been running everything stock, according to the CPU that is, so 200MHz HTT 1:1 memory = 200MHz (50MHz underclocked), set to Auto for most of the BIOS memory settings.
I had already tested my memory in my mobo, so I quickly tossed his memory into my mobo and started the built in memtest in the BIOS (Memtest+ v1.55)

I went for the memory Stock of 250MHz 3-4-4-8 2.6v.. I set my mobo's settings as close to this (http://www.xtremesystems.org/forums/showthread.php?t=95044) as possible, also lowering the CPU multiplier to keep it as close to stock as possible. Just like I had done with my memory.

Initial tests looked pretty good.. went to 260MHz 3-4-4-8 2.7v. Then I tried 265MHz and had errors at the very end of test 5 (same place my memory always errors out at).

So we let it sit at 260MHz for a while and started to mess with my memory in his DFI mobo.

OK same thing, matched his memory settings in BIOS to this (http://www.xtremesystems.org/forums/showthread.php?t=95044) exactly (since it's a DFI ;)), and again started at 250MHz 3-4-4-8 2.6v

Again same thing, but my memory actually went to 265MHz in his board and wasn't having any errors in test 5.

(Just for reference, we were manually setting test 5 to run by it's self. Our goal was to pass test 5 once, cuz the very last 2% of test 5 is where it always seems to error at. Once test 5 passed one run with out errors we would bump up the MHz and go again.. Once errors occurred we tried to bump voltage once, then twice to a max of 2.8v. When voltage didn't help we would backstep 5MHz, lower voltage again, and try for more than 1 pass of test 5.)

So I was like, "ok so is it the DFI that is making it run better?!?"

[kamakazibond]

I had heard that these sticks can run kinda hot, but with my setup and active cooling VERY close to the RAM my memory had never gotten much more than "slightly warm" in my mobo. So we deceided to see how hot it was in his case. Let me tell you they were HOT ! My friend was like "wow those are almost too hot to hold your finger on.." I was like, "they were never hot like that in my mobo.." The side of the case was off the whole time and I know that he has at least 2 case fans pulling air from the front.

So I believe that my memory did pass test 5 two or three times at 265MHz, and his memory did pass test 5 two or three times at 260MHz. Needless to say I was confused and he wanted to test his memory in his mobo.

So we did, we took my memory out and put his into his mobo, again starting at 250MHz and working our way up in the same manner. We also wanted to see how hot his memory got in his mobo.

I think we got his memory to 255MHz with out errors in test 5. we also felt his memory during test 5 to see how hot it was getting. Well during test 5 in his mobo his memory was HOT!, just as hot as mine had gotten in his mobo.

I had put my memory back into my mobo and was getting errors at 255MHz, I think. It also was no where close to as hot as it had been in his case. I deceided to try the other slots for my memory and see if they were any better. I had always used Orange and was attempting Yellow (yes my mobo is VERY similar to DFI's even the UV slots and connectors). Well the mobo didnt like that at all.. I ended up clearing CMOS at least twice while trying to get them working in yellow. Also switching one stick to slot 2 and the other to slot 4 and vise-versa.. So needless to say I went back to the Orange slots, but wait.. now I was getting errors at 250MHz!!! WTH?? Nothing had changed since I had been using my "CMOS Reload Program" (again VERY similar to DFI's "CMOS Reloaded"). But needless to say errors at 250MHz, no matter what voltage (2.6v - 2.8v).

So it appeared that his memory was doing better in my mobo and my memory was doing better in his mobo, but that all was about to change..

I tried his memory in my mobo at 250MHz... no go, still errors... now I'm getting pissed.. I think he still got to 260MHz with my memory with no errors in test 5. So he started a full test just for kicks, while I was trying to get 250MHz to work on my mobo..

Now at 260MHz with my mem in his mobo he got errors in test 3, so we tried voltage, then more voltage, nuthing.. still errors test 3, so we ended up at 255MHz with no errors in test 3 or test 5, so we let that run a bit.

But now I got a problem with my motherboard it wont do either kit at 250MHz anymore.. nuthing had changed but yet, no go..

[kamakazibond]

So after messing around with timings and whatnot, for no real reason I raised my CPU multiplier from 7x to 8x. Now all of a sudden, 250MHz with out errors. So I'm somewhat relieved and very, very confused. The whole time I was using the 7x multiplier and 2x HT multi, memory 1:1, etc.. now for some reason I cannot get even 250MHz with the 7x multiplier..

So now we have my memory in his mobo only at 255MHz and his memory in my mobo at 250MHz. So we again swapped memory, I put my kit back into my mobo and crossed my fingers with the 8x multiplier, and his memory in his mobo to see what we could do.

So yes my memory works ok in my mobo with CPU multi at 8x 250MHz 1:1, thank god, it seems stable again, but now I'm OCing my CPU (not really an issue, but we were trying to test memory and keep CPU as much at stock as possible.)

His memory in his mobo passed 255MHz ok and even into windows.. He ended up OCing his CPU some with a 10x multi and windows seemed ok.. Stable enough to run some Sandra benches anyway.

I think he still needs some stability tests run (prime95, superPI, windows memtest, etc), but I have run Prime95 for 8+ hours and Windows memtest for 4+ hours now on my system at 250MHz (8x CPU multi) and it seems stable enough, no errors or anything as of yet.

But here are my 2 main issues:

1st - Why have I now had 4 of these kits to test with, and none of them are good OCr's?? (asumming that it wasn't strictly a compatability issue with the MSI NF3mobo.) Isn't there tons of people with this memory well past 260MHz?? The noted info about them in the reviews and testing done by G.skill Tech says avg OC 270~280MHz.

2nd - I would not say that these memory perform any differently on a DFI mobo, compared to my own Nf4 based mobo anyway. It was a bit relieving to see his mobo have trouble running the memory as well, but on the other hand, it also means that the current batch (last 2 maybe considering my entire experience with a total of 4 kits) of GBHZ dosen't seem to be up to par with the previous batches.

I'll add as some notes that, we used the BIOS the DFI mobo shipped with, which may not be the most optimal one, but it had all the settings we needed to match up with this (http://www.xtremesystems.org/forums/showthread.php?t=95044) . Also I understand that our quick test 5 method may not be the most optimal, but from the windows based testing I did with my own mobo and my own GBHZ kit, I only got to 256MHz in windows using the method G.Skill Tech recommends. And I surely error at 256MHz in built in BIOS memtest. BTW both of our mobo's had the same version of memtest in it (memtest+ v1.55).

New extra note-- Yet another person today (or yesterday) has posted about his brand new from newegg GBHZ's not budging over 250MHz. More evidence that the current batch are of lower quality?? (http://www.xtremesystems.org/forums/showthread.php?t=96585)

[kamakazibond]

So it ends up that neither mobo and neither kit of GBHZ memory was any better than the other.

My next thing to do is test the GBZX's as high as I can get them and then prime, and sandra.. to compare with these loose timings at 250MHz with the GBHZ..

What is the RMA process if I were to deceide to keep the GBHZ's and RMA them for a better kit?? Do I send then to CA and then you send me new ones from CA?? I know of the RMA doc and emailing it to RMA@gskill.com or whatever, but the actual process.. where it goes who inspects and tests it, etc.

Like I mean will it just be shipped to TW and retested and put back into the distro?? Or will someone actually test it inside a mobo?

I still have more testing to do, just trying to research my options.

kamakazibond

[kamakazibond]

One last thing (I promise :))..

to somewhat backup G.Skill Tech's memtest+ weirdness.. When my friend wanted to OC his CPU with his memory up as high as possible, memtest+ passed without errors, and when trying to boot into Windows XP, he got a BSOD, So I do agree, that dos based memtest+ isn't fool proof anymore :(

[GSKILL TECH]

took me about 15min to finished all reading

something need to clear here

the one "codehunter" he bought at Dec 2005 not even 2006
RMA testing in done in the CA G.Skill lab with DFi ultra-d mobo & AMD 165 cpu
will run 260 3-4-4-8 for 2GBHZ, windows memtest, super pi 32m, and sp2004 at the same time for about 2~3 hours. if no error, we will increase to 265 3-4-4-8 and same tests again! if no error again. i will repick it and ship back to Taiwan for detail testing. no matter what, you will get a new one. the new for replacement, it will only test at 255 3-4-4-8 32M once. no error = ship em



basicly if you follow my bios setting and not overclock the CPU (around stock speed) you can easy find out the max for 2GBHZ. many kits same back i can easy to 265~280 if it is not defected. some of the kit become sample kit.

[kamakazibond]

something need to clear here

the one "codehunter" he bought at Dec 2005 not even 2006

Ahh yes thats when he bought it, but he just RMA'd to newegg and got a brand new replacement ;)


took me about 15min to finished all reading

And you said Give me a few days ;)..

Thank you for all of your help and attention, it truly is not normal for this kind of official support.


RMA testing in done in the CA G.Skill lab with DFi ultra-d mobo & AMD 165 cpu

:D I might be mistaken, but I was told DFI dosen't officially support Opty's :D


no matter what, you will get a new one. the new for replacement, it will only test at 255 3-4-4-8 32M once. no error = ship em

So on RMA's you do actually test on a mobo b4 u ship them out? I'm starting to like the sound of RMA to G.Skill over RMA to newegg, more and more.. I have seen what Newegg, NJ can do to a PC case during shipping 1st hand :(

And on second thought, I live in FL, so I get things from NJ alot, not 100% sure about this order, but A friend of mine with these GBHZ's has them at 300MHz, and he lives in CA. Since you hand deliver them to CA, maybe thats part of the problem, newegg's receiving in NJ. LOL they prolly throw and kick stuff around like crazy :D.


many kits same back i can easy to 265~280 if it is not defected. some of the kit become sample kit.

I might end up sending this to you just for the fact that I wanna see what u can do with it :), This has been the absolute worst experience I've ever had with memory. But I can say for one thing, that having you here to get info and help from, helps soothe the pain.

For the record, my friend is a bit disapointed as well, but I doubt he will RMA, cuz of the hassle, I don't blame him, but I've already gone thru soo much, whats another 2 weeks or so.. any estimated turn around rate, shipping wise.. I live in FL..

Anyway thats not a big concern of mine ATM, so get back to your other work, I'm gonna put the GBZX back in and try for 230MHz fully stable and see how the sandra bandwidth is.

Thankx for all your time and help, again!!

Amazing your are so responsive in the forums here, I hope G.Skill keep up the good work!!!

kamakazibond

[kamakazibond]

Well I tried to reboot, my PC which was stable at 250MHz for 3+ days, and it didn't POST.. :(

So I threw in the GBZX's and at 230MHz 2.5-3-2-5 I've got the best sandra memory scores yet.. Memtesting/sp2004 as I type this.. :)

So we'll see how this is going, IDK what to do at this point..

more, later..

kamakazibond