i would love some more voltage freedome. and MSI needs to look at it this way, if we kill their board with overvolting they get to sell another board!
Printable View
i would love some more voltage freedome. and MSI needs to look at it this way, if we kill their board with overvolting they get to sell another board!
My board has already been taken care of in the bios voltage "freedom" department. ;)
Okay guys I've been keeping this board on the side till they unlock the multi's so I haven't had too much time on it, but seems people are interested on 400+FSB so Here some SS I whipped up today, I kinda did a have ass job on mounting my cascade and didn't have enough time to redo the mount and contact so, All the results are of 10x to see what max FSB is and not max clocks.
Setup:
X6800 Retail @10x450
4x 512MB 8000UL Fattys 1:1 4-x-x-x
MSI 975 X V2 715B Stock HS with fan
7950GX2 Stock
2x80GB Seagate 3.0 + 1x74raptor in Raid 0
PC P&C 1KW
Vcore 1.7v modded
Vdimm 2.75v. modded (the mods are here (kyosen's I think), do a search, if you can't find it PM me)
Pcie 1.8v (set in Bios)
1. 450x10 4500Mhz SPi 1M
http://img218.imageshack.us/img218/4...0fsbaj8.th.jpg
2. SPi 2M
http://img226.imageshack.us/img226/9...pi2mck5.th.jpg
3. SPi 8M
http://img226.imageshack.us/img226/2...sb8maw2.th.jpg
4. Everest Mem Test
http://img228.imageshack.us/img228/9...restac0.th.jpg
5. Cinebench
http://img228.imageshack.us/img228/1...enchnl6.th.jpg
6. Sandra Bandwidth Buffered
http://img244.imageshack.us/img244/2...bfmbhh0.th.jpg
7. Sandra Unbuffered
http://img244.imageshack.us/img244/8...bfmbej8.th.jpg
8. Sandra Cache (This I guess tells you how the boards strapping the chipset)
http://img225.imageshack.us/img225/6...achemv2.th.jpg
9. Aquamark with 7950GX2 stock HS clocked @ 600-800
http://img151.imageshack.us/img151/1...212ktv0.th.jpg
10. Now all just SPi 1M to see max fsb 460x10
http://img151.imageshack.us/img151/9...0fsbsa1.th.jpg
11. Finally the Max FSB without vmch mod or phase on chipset, I think some more left, but wouldn't be worth the time without mods & cooling
http://img151.imageshack.us/img151/4...b467xj9.th.jpg
Just a note: Well I notice there's a new bios so this may not be a problem anymore but here goes anyway, I'm thinking the boards problem with booting and sticking the bios setting has alot to do with MSI Dual cell bios recovery and chipset strapping, if you guys notice most 975X board do a full shutdown to reboot after changing to a high clock FSB or changing the straps on the chipset, but I think MSI can't do that correctly because if you ever owned a MSI with Dual Cell Bios, you'll know hitting the reset 2times will reset the Overclock to default. Every time it get stuck after a settings change, the board hangs in code 26 of my debug card. I notice thats where the voltages change during a successful boot. So, I figure, if it reboots to change settings it will trigger the Dual Cell to reset, so it just hangs? I found that I just have to wait abit and let it hang for 20-30 secs then I hit the reset myself, about 98% of the time it reboot with the setting I wanted (only hit it once or if you hit it twice you'll be reset to default). Well thats my 2Cents,but I maybe completely off and barking up the wrong tree:D
Wow nice clocks and thanks for all the great info :).
sson74, why is your vdimm set so high? I would have thought that with your high spec ram you could run at default voltages and easily reach 1:1 even with an FSB of 467!
System specs OCZ Gold XTC x 2GB, E6400 BIOS 716
Test 1 FSB 320, Ram Divider 1:1.25 (=DDR800), VDIMM 2.15V
passes memtest no problem
Test 2 FSB 375, Ram Divider 1:1 (DDR 750), VDIMM 2.15
No Post (won’t post past 350)
Test 3 FSB 375, Ram Divider 1:1 (DDR 750), VDIMM 2.2
Post but errors on Memtest
Test 4 FSB 375, Ram Divider 1:1 (DDR 750), VDIMM 2.25
Passes Memtest 1 pass tested so far
Test 5 FSB 400 Ram Divider 1:1 (DDR 800) VDIMM 2.1
Posts no problem. passes memtest no problem
I sent back my first pair of OCZ memory thinking the memory was defective, as it failed memtest. Also tried the same test on GSkill F2 6400 PHU2 1GBNR ram, same results need to overvolt the ram is running at 1:1 divider but not if I run FSB 320 with 1:1.25 divider.
Which begs the question why do I need more voltage when the divider is 1:1 at DDR 375 versus Test 1 where the Ram is running DDR800 at stock VDIMM by using a divider? Poor Voltage regulation on this motherboard? This seems to agree with the results by sson74 as he needs alot of VDIMM to reach high FSB.
I'll post this at the MSI forums to see if I get any good answers.
i had it like that from before, and, i just wanted to run some numbers really quick for fsb, didn't bother to change anything from my prior session.Quote:
Originally Posted by idioms
But I must say running 1:1 @450+ needs more volts than @ 4:5.
Edit: here's a SS at 11x431 with tighter timings and higher clocks yet look at the BW and compare to the 450x10 @4-x-x-x everest above
http://img227.imageshack.us/img227/1...0mhzgj6.th.jpg
What effect does raising the PCIe voltage have on the graphics card? Are any of its internal voltages changed by this?
Well, I haven't notice anything on this board but I keep the speed @ 100 (I assume you know PCIE volts=Vmch).Quote:
Originally Posted by idioms
This is a good read about vmch(pcie) volt
first observations
1. The cold boot issue is better but still not fixed. for clocks higher than 360 I still get the issue.
2. Reboot still doesnt work
3. No improved overclocking
4. No additional options
Overall, it "feels" a little more stable so its worth updating especially if you are clocked under 360 FSB. A couple of times at FSB 360 the bios hung on me, this didnt happen at 333, 380 or 400 i would be interested to see if anyone else sees this behaviour
I think something happens in bios at FSB 360, maybe a change in chipset strap .. the board seems to consistently have issues at this exact point (FSB 358 works perfectly and consistently)
There is definitely something going wrong in dual corecell, I had to reset after upping the FSB to 380 (back to 266), so i started increasing FSB again, set it at 333 .. I then shutdown, started up and up came FSB 380!
I wonder if there is a way to "disable" dual corecell? or force it to use a single "cell"
update
This is spot on. those of you that are having trouble getting to FSB 400 should try this. I find it quite difficult to consistently get this board over the 360 hump but this piece of advice has improved my hit rate considerably .. you have to wait at least 20 seconds .. perhaps something is being replicated between the two core cells? you then get a warm boot from reset and this seems to help with the higher FSB.Quote:
So, I figure, if it reboots to change settings it will trigger the Dual Cell to reset, so it just hangs? I found that I just have to wait abit and let it hang for 20-30 secs then I hit the reset myself, about 98% of the time it reboot with the setting I wanted
On the downside i still get the cold boot issue
could someone post a vcore mod in an easy n00b manner for me?
ok guys, when I overclock to anything past 370fsb, the pc boots to windows. I've got it to boot to windows @ 3.9Ghz, but when put underload, it reboots. Is this the reboot issue you guys are speakin of? I maxed out voltages on everythin, vcore, ram, pci e. what could it be. my E6700 is lapped and watercooled. temps are 57-61 underload at 1.52 vcore and @ 3.7-3.8 Ghz.
Nope. For me this board doesnt reboot from windows when FSB is set past 360, it just hangs. (interestingly it does reboot at FSB less than 360 .. i would be interested to know if you guys with higher multi's get this problem earlier than FSB 360?)Quote:
Is this the reboot issue you guys are speakin of
for you hotrodtheman i'm almost certain you have reached the highest speed your cpu will go at vcore 1.58 (you meant to write this right? 1.58 is the max this board can do .. although it probably is 1.52 after vdroop) .. welcome to the wonderful world of volt modding.
Lol. I hear ya on the volt modding, and maybe putting money aside for a vapochill... im serious. I had it maxed out at 1.58v @ 3.8ghz and 3.9ghz and it still rebooted under load. I was giving temps of the core at 1.52 which is what it runs at 24/7 (1.48 after vdroop I believe) Anyways, good luck guys. I'm happy with a 1ghz overclock for now. I'm gonna keep an eye on coolalers thread and see how that DFI is working out. Hopefully in the near future money will be better and I can change boards :)
http://www.nickfire.com/msibios/975x...p/7246v716.zip 716 BIOS, re-hosted. :)
Not sure with this statement. For me if I change FSB settings more than say 30 the board will hang after saving to cmos. When it hangs I just hold the power button for 3 sec, board shuts down then I fire her up and she boots fine with last settings every time.Quote:
So, I figure, if it reboots to change settings it will trigger the Dual Cell to reset, so it just hangs? I found that I just have to wait abit and let it hang for 20-30 secs then I hit the reset myself, about 98% of the time it reboot with the setting I wanted
MY problem is no matter what FSb I am at if I run anything but 1:1 memory I will get a cold boot hang. The rig will pass everything i throw at it running up to 1000mhz memory speed and will even restart fine as well. It's when I shut down the rig, go get a soda or even go to bed, when I go to power up again the board hangs and will not go into bios untill 3 restarts at which time bios is set back to 266fsb:mad:
I just do not understand how this board can run my CPU at 3.7ghz and 1000mhz memory, crushing everything I throw at it in Windows and then the next morning refuse to boot:stick:
Maybe there is a cirtain way we are supposed to save to cmos for the bios to save and hold and then be happy with the new settings.
rodman, I'm interested to see if you can still run memory at 1000mhz with any bios besides the shipped 7.0. I could also run my ram at those speeds but neither of the new beta's will allow me to.
I really see no reason for most to flash to any other bios as 7 does the job just fine. I would flash back if I were you Racer. Untill an official bios comes out that fixes the CPU ratio select and running memory over 1:1 I see no reason to flash the bios.
Who really cares if bios does not report cpu speed correctly in bios, Windows and benchies tell me 3.8 is 3.8ghz, futhermore with my setup this board is very fast, faster than anything I have tested;)
Oh, yae. I did some testing today running diff memory dividers, namely 1:1 and 1:1.25. I ran the memory@902mhz@4-4-4-10 @3.6ghz or 360fsb. Then I ran ran memory at720mhz@3-3-3-8@3.6ghz or 360fsb. I used Aquamark as a test along with 3Dmark and found the scores to be within 100 points of each other.
I find this interesting since running the memory at a higher bandwidth really did not provide any more performance in Gaming situations. It really felt like I had an A64 with results like this. I will say that running 902 mhz did result in higher benchmarks but not by much at all, nothing a gamer would notice.
So with that said I don't see the need for running high bandwidth on a conroe, rather tight timings and lower speeds will give you plenty of performance unless you need every point in a benchmark. Infact I don't really need the fix for the memry divider on this board as it really won't help me that much.
Try it out for your self and see:)
I got some connections @MSI and right now they're testing the "cold/warm boot" issues, memory compatibility issues and thinking on unhiding the multiplier option ;)
lads....... what do i do?....my msi 975x power up has arrived, but i also have a p5b deluxe on order...................i have a "cool off" rma for the 975x as long as i dont open it.....but........what do i do....im dying to give it a shot.......
i know the p5b deluxe is definatley the higher fsb option and i have a E6400 e.s. and an E600 retail sitting....help me chose lads......
Gordy!:confused:
I'd go with the MSI. It can do 400FSB, we know that. Stable, in most cases too. That's more than enough for your E6600. TodB claims he has connections at MSI (wicked sweet :D) and says they're working on the boot issues. So yeah, I'd go with the MSI.
I can't get up to 400 FSB, at 390 its rock stable, but at 400 it fails. Tried everyting, I hope not, and it just wount be stable at 400.
E6600
Ocz 6400
vcore 1.56 - will not help to raise it to 1.58
pciev 1.65 - haven't touched
vdimm 2.2 - tried 2.1
1:1 5-5-5-15
Anyone got any ideas?
Watercooling, E6600 load at 54C max, watercooling at NB too.
Up your PCI-E voltage a notch or two and you'll prolly get 400Mhz stable ;)
PCI-E voltage = Northbridge voltage (MCH)
I tried now, even at 1.75 pciev 400 isn't stable, but at 390 it's stable at even 1.55 pciev.Quote:
Originally Posted by NickS