Thanks for the tip, I have been trying to run +3 on one and stock on two. I'm still reading that thread so shhhhh.... :p:
Printable View
Thanks for the tip, I have been trying to run +3 on one and stock on two. I'm still reading that thread so shhhhh.... :p:
look at this giga ddr2 P45 chipset board that will come out soon...GEEZ!
CREDIT:... http://www.ocworkbench.com/
Quote:
Originally Posted by OCW
The heatpipe designs are getting more elaborate all the time. Nice layout I think, I like the slot config.
yeah!...this is what stands out, which peaked my curiosity:
Quote:
Outstanding cooling design with Pure Copper thermal design featuring Water Cooling
Is that kit F3-12800CL7D-4BGHZ? If so would you mind sharing setup, timings and anything else useful? I'm currently using them at 1800 with 8-8-8-20 2T 1.95v all others auto. Is there something better I can do with these?Quote:
Originally Posted by Ace-a-Rue
:D are you kidding me...sheez!;)..i never got up that high even with my E8500....i guess my sticks are doing sh*t for me!:p:
this is what i did with my 9650:
http://www.xtremesystems.org/forums/...&postcount=120
EDIT:...here is my best with my E8500 (800 mhz): http://www.xtremesystems.org/forums/...&postcount=233
Ok so I'm not totally on the wrong track... good to know. Thing is a 2x1gb ocz flex kit really can't produce much better results. Maybe it's the cpu, I had a chance to trade for a 9550 and now I'm wondering.
no, no!..2 x 1gb kit will do better than 2 x 2gb kit, just your sticks (we have the same) are doing better than mine.:up:Quote:
Originally Posted by loonym
Testbed:
CPU: E8400
Memory: Gskill GBPI PC12800 (1600) 2gb x 2 kit
Mobo: Gigabyte X48T DQ6 Rev 1.1 Bios (F4h)
Gpu: Gainward Bliss 9600GT 512mb Golden Sample
HDD: IDE Hitachi 80gb 7200rpm
PSU: Vantec Ion2 500 watt
OS: XP Pro SP2
Memory is 7-7-7-18@ 2T@1.9vdimm rated by manufacturer
Of course, I am not going to run at manufacturer rated clock speed
900mhz@8-8-8-8-18@2.1vdimm - Pi 32 million passed
http://img232.imageshack.us/img232/4...881821vgh9.jpg
900mhz@8-8-8-8-18@2.1vdimm-01se
http://img215.imageshack.us/img215/4...881821vrc0.jpg
900mhz@8-8-8-8-18@2.1vdimm-03
http://img238.imageshack.us/img238/7...881821vnl3.jpg
900mhz@8-8-8-8-18@2.1vdimm-05
http://img89.imageshack.us/img89/658...881821vlv6.jpg
900mhz@8-8-8-8-18@2.1vdimm-06
http://img171.imageshack.us/img171/3...881821vmv7.jpg
So far, Bios F4h is the best bios for memory oc. Bios F5a and F5d failed to get the memory stable at 900mhz=(.
Here is the memory module
http://i234.photobucket.com/albums/e...w/SANY0134.jpg
I've been having good results and stability with bios F4D so I really haven't tried others for more than just a look. All my benches as well as 24/7 operation running boinc have been pretty much with the settings I posted. I'm never going to get any higher with this or any other board and the x3360's because they won't run past fsb 475 or so.
pretty well 475 fsb is it for quads on air...i think i got to 480 fsb but that was pushing it with voltages...a bit beyond my comfort zone;)
Yes I agree and with all the reading I do I guess I knew it even though I wanted to prove it wrong, lol. It's so much fun to try and beat the odds you know? :DQuote:
Originally Posted by Ace-a-Rue
I have had this plan in the back of my mind for awhile... There's a local tech school that has hvac/hrac courses and a pretty nice shop. My thought is to approach 1 or 2 of the students with the idea of building a really nice 1hp or so ss for me as a school project with me funding the neccessary supplies. Never hurts to try :)
Hi folks,
I'm working on another machine for my office this time. GA-X48T-DQ6 with a QX9770, Quad 15k SAS Enterprise and a Promise SuperTrax 8650 SAS controller; also 8GB of DDR3-1600 memory from OCZ. Now I have the memory passing at least 1 pass and still running no errors, 4x2gb at 8-8-8-25-2T with 900mV CPU drive and "Normal" or "Auto" skew.
I am using BIOS F5C which I found with google. Under BIOS F4 I couldn't even get 1 stick of memory to pass memtest with DDR2-1600 settings, albeit I was trying at 7-7-7-24-2T 1.9V, you know the rated specs for this memory (should have worked, right)? Anyway...
When I go into the BIOS on F5C and try to enable the FSB control, it quickly disables it again. I have absolutely no idea why. I want to try the memory at DDR3-1800 with FSB 450 but maybe this BIOS has a bug in it, can anybody else comment on this? Thanks.
(PS. In my signature rig below I am swapping out the 750W Power for a 1000W power because my Kill-A-Watt recorded 1010W in 3dMark06 peak so don't razz me about that!)
BIOS F5C is flawed for the reason you found. Use F5A or F5D
luck:)
The link to station-drivers has no file found so where can I get F5D now?
Link works for me
see * GA-X48T-DQ6 BIOS v.F5d
It's in the 2nd box down
Quote:
Gigabyte company released new Bios version for its motherboards
In new version some errors are fixed and work improved
To download bios update choose your motherboard model from following:
* Gigabyte
o GA-X48T-DQ6 BIOS v.F5d
o GA-X48-DQ6 BIOS v.F7g
o GA-EP35-DS3P BIOS v.F4e
This is very strange I cannot get to it:
http://www.station-drivers.com/telec...os/x48tdq6.f5d
Traceroute dies here:
9 17 ms 20 ms 15 ms 68.85.162.58
10 24 ms 35 ms 20 ms 68.85.162.57
11 49 ms 48 ms 50 ms te-0-10-0-5-cr01.newyork.ny.ibone.comcast.net [6
8.86.90.45]
12 48 ms 51 ms 49 ms tengigabitethernet1-4.ar5.NYC1.gblx.net [64.208.
222.57]
13 123 ms 125 ms 125 ms 020G.gsw-1-6k.routers.ovh.net [213.186.32.130]
14 138 ms 129 ms 149 ms 020G.rbx-1-6k.routers.ovh.net [213.186.32.226]
15 129 ms 131 ms 128 ms rbx-34-m1.routers.ovh.net [213.251.191.97]
16 *
Can you send me the F5d file? jking _AT_ prospeed _DOT_ net
Thanks.
MagnumMan u have a PM:)
good job Fugger sorted us out :lol:
Well with BIOS F5D and a little bit of time I was able to get 4x2gb (2 kits of OCZ3P16004GK) running stable under memtest.
FSB 410
Mem DDR3-1640 +.40V = 1.92V
Timings: 8-7-7-20-2T
All other settings on auto in timings
FSB +.20V
NB +.20V
CPU x9 (QX9770) = 3.7GHz (not tuned yet, I hope to break 4GHz...)
CPU at 1.35V right now
CPU Driving 900mV (normally 800mV)
No skew, left on "normal".
Not too bad for 4x2gb dual rank memory these days. The memory does get pretty hot though I would almost recommend a fan right on the memory to help out. They almost burn to the touch when pushed.
I mounted a 40x40x20 fan on the northbridge so it has active cooling (I left the heat pipe solution alone and put a fan over it securely, plugged into the "NB" fan header). The NB is almost warm to the touch near the base when running memtest. Before the fan it was getting almost too hot to touch.
Anyway so far so good. I tried 8-7-7-18 but the system kept boot-looping. So this looks like the best timing I can get with 4 sticks. Good board, seems pretty stable.
my cold start problem went away
it was the damn power supply
p.o.s. Seasonic X900 :down:
Corsair HX620 went in and all is sweeeeeet :D
I want those rams :DQuote:
Originally Posted by dinos22
good news!...which bios are you using?Quote:
Originally Posted by dinos22
I finally stabilized a rig with this board. Took more reboots than I thought it would. Here are all the specifications to help people out who may be setting up something similar.
QX9770 Rig Settings and Benchmarks
----------------------------------
Mobo: Gigabyte GA-X48T-DQ6 BIOS F5D
Custom active cooling on northbridge heatsink
CPU : Intel QX9770 @ 3.8GHz
Scythe Mugen Cooler (it's huge)
Mem : 2x OCZ3P16004GK (4x2gb PC12800)
Disk: 4x Seagate Cheetah 15K.5 SAS 146GB 15k
Ctrl: Promise SuperTrak EX8650 (800MHz ioproc)
Vid : ATI Radeon 3870 @ 825/1150
OS : Windows 2008 Server x64 Enterprise
BIOS Settings:
CPU Ratio 9.5x => 3.8GHz
FSB Control => Disabled (400MHz)
PCIe Frequency => 110
CIA2 => Disabled
Perf. Enhance => Standard
Memory Multiplier => 4.00D
Memory Frequency => DDR3-1600
Memory Timings => Manual 8-7-7-20
tCL 8 (CAS)
tRCD 7
tRP 7
tRAS 20
Secondary Timings
tRRD 5
tWTR 6
tWR 12
tRFC 75
tRTP 6
tRD 9
Adj 2
CMD 2T
CPU Drive => 900mV
CPU Skew => Normal
MCH Skew => Normal
Voltage => Manual
DDR3 => +.35 (=> 1.88V)
PCIe => +.10
FSB => +.20
MCH => +.35
Load => Enable
CPU => 1.38750V
Benchmark Scores:
SuperPI 2M 29.437s
Sandra XII SP2c
Processor Arithmetic 49990 MFLOPS 64218 MIPS
Inter-Core Bandwidth 22202 MB/s
Physical Disk 251.84 MB/s @ 6ms
Memory Bandwidth 7.91GB/s (int/float same)
Memory Latency 74ns (92.40 speed factor) L1 3 clocks L2 18 clocks
Cache + Memory Bench 59.06 GB/s (56.50 speed factor) L1 321.66 GB/s L2 101.01 GB/s
Everest 4.50.1330
Disk Read
Linear Begin 377.3 MB/s
Mid 335.3 MB/s
End 227.2 MB/s
Random 320.4 MB/s
Buffered 1045.6 MB/s
Access 5.7 ms
Memory Read 8738 MB/s Write 8457 MB/s Copy 8691 MB/s Latency 64ns
L1 Read 60708 MB/s Write 60556 MB/s Copy 121097 MB/s Latency 0.8ns
L2 Read 27559 MB/s Write 20219 MB/s Copy 29482 MB/s Latency 4.5ns
HD Tune 2.55
Burst 397.9 MB/s
Avg 153.0 MB/s
Max 172.6 MB/s
3dMark Vantage 1.0.1.1 P4933 (4024 GPU, 15319 CPU)
can you get to 4 ghz and run any benches?
Unfortunately Core #1 (numbered 0 through 3) dies above 3.8GHz at any voltage and quickly too. Since temps are already 80C (175F) !! I prefer to keep that where it is. Took me 5 days to stabilize this rig, I'm done doing that. This is a development rig for my office which I will be using to build an app that takes over an hour to compile/link on an Opteron 248 (2.2GHz) system. I expect that time to be around 20 minutes with this new hardware. Keys to stabilizing were:
1. running memtest to find the memory stable settings (DDR3-1640 8-7-7-20-2T), anything over that was failing frequency-wise.
2. running prime95 small FFT to isolate the CPU's potential, which seems to be 3.8GHz (this is pretty lame because other people have gotten to 4.0 GHz, so I think I got a bad CPU, but it's still well within engineering tolerance, I suspect).
3. upping voltage on the FSB and MCH, FSB at +.30V locked up the system; MCH at +.40 locked up the system.
4. setting the CPU drive to 900mV
5. Manually setting all the secondary memory timings to be one higher than the auto selection was, i.e. 5 became 6.
The Scythe Mugen heatsink is so large it rests on coils too, which may be preventing it from getting the tightest fit possible on the CPU... 80C seems pretty excessive even at 1.38750V to me. Then again I am used to water cooling (see sig) and this doesn't have it.
Nice setup MagnumMan. It sure seems like you're right when you say the heatsink might not be making full contact. Keep an eye on those temps. I've started using 100 for cpu skew and I think it's helping stabilize my highest overclock attempts. I also use the GTL settings that Supershanks was kind enough to give me and that made a difference too. If you press Ctrl F1 at the main bios screen you'll see those GTL settings available when you go in MIT section.
i got it!...you are P95 limited...that bench is a stopper for a lot of people trying to maximize the overclock.Quote:
Originally Posted by MagnumMan
the high temps, i presume, are from the P95 runs..right?...if so, they might be about 7-8C higher than i would expect from a 3.8 ghz overclock which was P95 capable...what kind of thermal paste did you use and what is the fan speed set to?
Thermal paste: Arctic Silver V
Fan: BIOS driven, I think it runs 1200rpm (120mm fan) + the rear exhaust fan which is about 2 inches from the heatsink is a new Scythe S-FLEX SFF21F 64cfm fan.
I do not like the SCINF-1000 Mugen's locking clips, I don't have a good feeling they are "secure". I would have preferred a backplate with screws. It came with a Air Flow: 46.5CFM fan.
My guess is a very thin copper plate would be best to get the cooler from sitting on the choke coil blocks too. It looks like a good fit, but my guess is it is marginal - working, but not as good as it could be. The bottom of the SCINF-1000 is huge for compatibility with lots of CPUs.
Hi guys
any reports about issues with this mobo + OCZ PC3-14400 PLATINUM ? Or issues using the 4 slots?
The default SPD Vdimm settles at 1.728v, which sucks.
Every tiny increase in Vdimm = lots and lots of memtest errors and boot lockups. So much for the 1.95v on the sticker. I can't even go past 1.8v.
considering this is all at 1333mhz (pretty good for 1800mhz mem and x48t, right? right? :shakes: ), I'm off to a greaaat start with this combo.
edit: F5D
have you tried F5A bios?
i would recommend the TRUE 120 heatsink and then a panaflo 38MM fan.Quote:
Originally Posted by MagnumMan
i used to use AS5 but i transitioned to MX2 paste...it might be a tiny bit better and WAY easier to apply.
See my QX97770 post above. You cannot rely on Auto settings for this to work especially with 4x memory sticks. You didn't mention if they are 1gb or 2gb, single or dual rank memory.Quote:
Originally Posted by |Oc|REvO_HsArC
You will need to set the Vdimm to +.30V => 1.88V or +.35V => 1.93V. I was able to use +.30V. You will need to set the northbridge (MCH) voltage much higher than stock, I am using +.30V on the northbridge, note though I slapped an active cooler fan on it, 40x40x20 about 9CFM quiet, with some hollow .5" spacers and some 2" drywall screws going right into the NB heatsink fins (between them). You will probably need to start with 9-9-9-30-2T and all the secondary memory timings set to be +1 over the Auto value, except for the one that defaults to 60, try 75 or 80 on that. Good luck, it took me about 5 days to stabilize my QX9770 rig, and seriously countless reboots.
First I would run memtest on the memory and get it stable. Then push it to the highest speed / lowest latencies you can stable. This will take a couple days... second use Prime95 Small FFT to isolate the CPU and make sure it'll run at the FSB, Vfsb, Vcore, Multiplier. Finally, once those pass, run Prime95 "Blend" and give each instance of the app a different CPU to run on (it is in one of the options) and give each instance 1/4 of the available free memory in the "Custom" section.
If you can get all of these to run 6+ hours straight I would say you have it stable... the last one should run 24+ hours without a hitch as it tests everything.
Ryder at OCZ support (on the forums) is responsive and has one of these boards for testing. He suggested leaving all the secondary timings at Auto... I found they had to be tweaked in order to stabilize the platform.
I'm about to. I was delaying it because F4 and F5D behave exactly the same, making me think its something else
Will try that (well I did, kinda, FSB overvoltage at max, but mch only a bit above)
what about your sub-voltage support like MCH?Quote:
Originally Posted by |Oc|REvO_HsArC
yes I'm about to try bigger values. But Its really weird that everything is stable at 1333mhz 1.78vdimm, but at 1333mhz >1.78vdimm is very unstable. Like the classic syndrome of mems not liking voltage.
It's true to some extent that the mem will puke on the voltage until you start getting more agressive with the fsb as well as timings and other board voltages like Ace said. I wouldn't have thought this until I read the 'safe micron DDR3 voltage' thread in the bw section. I learned tons from it.
that link is interesting;)...many have their opinions:D...i kept my ddr3 mem at 2.05 and below....i haven't found the ddr3 read point on this board..have you?
i wasn't the biosQuote:
Originally Posted by Ace-a-Rue
it was the power supply 100%
also i am now on F5a
good bios
was benching some sick speeds with this board....really like the board a lot
have a look at these RAM timing
this mobo is a crazy MOFO :D i tell ya
http://img296.imageshack.us/img296/9...4200mhztd7.png
if you guys are trying to get my FSB out of the board try using 150ps CPU Clock Skew...........it should help at least on dual cores...maybe quads too
Yep I lost a set of cellshock on the Asus P5E3 Deluxe went exactly as described, had Corsair dominator memory cooler on and never went over rated volts, was still learning :lol:Nice going Dino:up:
Gigabyte technical support, on the other hand, takes about twice as long as MSI to reply and when they do it wasn't very helpful... the Gigabyte technical support staff is not integrated with their Engineering staff, which is rather annoying. MSI technical support could supply the latest BIOS and even get a custom feature into a BIOS like PCIe frequency select on the K9A2 Platinum. So I'll score Gigabyte low on the technical support, but very high for a very stable motherboard. I went through 4 MSI K9A2 Platinum (AMD) boards before I got a stable one, I only bought one GA-X48T-DQ6!
Answer - 616325
Answer : We do not have it on our database, we will need to check with our bios team for availability
Question - 616325
From : ME
Sent : 6/10/2008 09:21
Question : Are you telling me you are unaware of the Beta BIOSes F5a, F5c, F5d out there? That doesn't make sense...
Answer - 615271 (on 6/10/2008 around 1am)
Answer : Where did you get this bios from?
According to our website only F4 bios is available
Question - 615271
From : ME
Sent : 6/7/2008 03:09
Question : I am looking for a newer BIOS than F5c. I have F5c stable on 4x2gb DDR3-1600 memory and I want to move the FSB to 450, but I cannot because in BIOS F5c the FSB selector is always disabled. Do you have a F5d or F5e you can send me?
-------------------------------------------------------------------------------
Model Name : GA-X48T-DQ6
--------------------------
M/B Rev : 1.0
BIOS Ver : F5c
Serial No. :
Purchase Dealer : newegg
-------------------------------------------------------------------------------
VGA Brand : ATi Model : 3870
CPU Brand : Intel Model : QX9770 Speed : 3.2
Operation System : Win2k8 x64 Enterprise SP :
Memory Brand : OCZ Type : DDR3
Memory Size : 4x2gb Speed : DDR3-1600
Power Supply : 750 W
at least you are getting some feedback although it sounds like the right hand does not know what the left hand is doing...hahahahahaha:DQuote:
Originally Posted by MagnumMan
agree with a good stable board!:up:
Quote:
Originally Posted by dinos22
that is a damn good 32M SuperPi!!!:up:
|Oc|REvO_HsArC, which strap I set makes a huge difference too in what I can do with my mem. My best luck on the F5D bios has been with 400b.
This is one...Quote:
Originally Posted by Ace-a-Rue
I'm not to that point yet.Quote:
Originally Posted by loonym
First I need voltage, and anything above 1.78vdimm = thousands of errors in memtest, at stock speeds (1333mhz). These mems are rated at 1800mhz 1.95v, so this is a problem.
edit: after 38h of memtest:
1333mhz 1.72v rock solid.
1333mhz 1.8v Thousands of errors
see my point?
I think this board was a mistake. I don't have a voltmeter, but i'm pretty sure when I "increase" the vdimm, it actually decreases.
just now I increased 0.10v over the default 1.728, and it reads 1.680v. I increase another 0.05v and it reads 1.775v... lol
bios tested: f4 f5A f5D
:lol: It overvolts slightly I've run my cellshock at +0.35 24/7 giving 1.85v this generally appears on everest & bios as 1.90vQuote:
think this board was a mistake. I don't have a voltmeter, but i'm pretty sure when I "increase" the vdimm, it actually decreases.
Interestingly enough i'm waiting for a replavement set after a stick died on me last week., exactly as described, no fade out , dead. it was the 2GB CellShock DDR3 PC3-14400 1800MHz (8-7-6-21) CS3222580
Dealer confirmed yesterday they were out of stock so is repacing them with 2GB CellShock DDR3 PC3-15000 1866MHz (8-8-8-16) It is a bit disturbing as this wil be my 3rd set. My 1st went back after memtest errors on Asus P5E3 Deluxe Wifi
I've had a temp sensor on the inner side of a module connected to my mcubed and tempos have been cool . The heatsinks have always felt pretty cool. I've continued the habit fronm ddr2 days of running with a Corsair dominator memory cooler . so:shrug: I've not been a cruel master :rofl:
Well then my guess would be this board really fails to read my mem spd.
either way I'm stuck with this setup because DDR3 is new and I don't have any different components to test either thing.
at least ddr1300 is stable with that spd voltage.
thanks once again.
bye
can you tell me where that is on the board?Quote:
Originally Posted by loonym
EDIT:..ok, i found the location...i will give it a try.
Aye X marks the spot matey :rofl: Zoomed in out & around some of my pics to find it :)Quote:
EDIT:..ok, i found the location...i will give it a try.
Thanks loonym :up:
i endorse that one!:up:Quote:
Originally Posted by Supershanks
The board seems to regulate the vdimm quite well I think. With bios F5D and the 4gb gskill kit set at 1.95v and running ddr3-1800 8-8-7-20 2T stressing with boinc and hyper-pi simultaneously the reading from that point behind the ram slot is actual 1.944v or as close to spot on as any board I've seen and certainly light years better than the p5e3-premium I ditched.
yes I had the x38 version the deluxe & wasn't that impressed with ram voltageQuote:
I've seen and certainly light years better than the p5e3-premium I ditched.
I'm going to check mine tomorrow. I've got a strange case of deja vu with this, Think I did it when I first got the board., I poked around with my MM. I remember being a bit wary after Ace's attempts. :shrug:
:rofl: It's actually an ideal spot. I can steady the probe against the 4th ram slot and check it no troubles even as bad as I shake :D
No the vdimm is easy , it's vcore that's the hard 1 i think on this board. Ace did a pic of it somewhere.
Found it :)
here is the Vcore measuring point!
Oh and
BUMMER...board died on me!..Ouch :eek:
Really need to avoid the shakes there.
Yeah there's another vcore point that's just as scary on one of the legs of that ic just at the bottom of his photo I think. I inadvertantly shorted somewhere in there yesterday but luckily only got a reboot. :shocked:
Yep I shorted mine a couple of times , just probing, but have been lucky :)
LOL...that short was due to carelessness on my part and also not having electrical tape around the probe almost all the way down to the point...i now have about 1/4" or 3/8" of the probe point exposed...now, if i put the probe at to much of a shallow angle, i should be protected against grounding it on any other electronic part.
:lol: I have visions of sneezing, whilst using the mm :ROTF:
Not recommended for hay fever sufferers :shrug:
YUP!...i have sneezed while measuring!:DQuote:
Originally Posted by Supershanks
I'm not gonna throw rockets yet, but my issue with the voltage - as odd as it seems - seems to be related to the Command rate. I would never guess it due to the nature of the problem.
1333mhz 1.72v 1T = OK
1333mhz >1.72v 1T = NO-GO
1333mhz >1.72v 2T = OK so far
i don't know if it so much the vdimm voltage that it the difference between 1T and 2T...you might be right on the border of having safe (usable) memory settings to function at 1T
EDIT:...have you increased the CAS & PL setting up by one notch when you set 1T with a higher vdimm than 1.72v?
i didn't set 1T... it was there by SPD and I never touched it. I think I tried every usable cas latencies but always failed with higher voltage.
So far I'm doing 41minutes of memtest 1600mhz 8-8-8-24 2T 1.92v. It was definitely the command rate.
lol how absurd is that? 1T failing with higher voltages? shouldn't be the other way around? Boy things have changed since my overcloker days back in sdr and ddr1 times...
Dusted off the multimeter today & ran a few tests
Voltage reading points & testing
I've found a couple more voltages FSB & CPU Clock Driving. Control
I think i've found (G)MCh but there are some oddballs.
Would appreciate your views :D
Very nice stuff Supershanks! Thanks :up:
that is a problem with this board...it does not handle SPD settings very well...it is way better setting them manually.Quote:
Originally Posted by |Oc|REvO_HsArC
CORRECTION:...the positive probe point for my DMM has less than 1/8 of an inch of it exposed to avoid those accidental shorts that might kill the board.Quote:
Originally Posted by Ace-a-Rue
YES!...i agree with loonym!...i answered you in your thread about the MCH voltage..i am thinking it is 1.25v as the default voltage for MCH.Quote:
Originally Posted by Supershanks
I haven't really read the anand article. I suppose they didn't indicate read points, I further suppose they posted whatever number gigabyte fed them in the review. It would seem strange to me to see such solid voltage regulation everywhere except that one value and the readings I came up with pretty much matched what you indicated Supershanks. Thanks once again.
Your Right :)Quote:
I suppose they didn't indicate read points
Update:-
Tested FSB with wolfdale, can confirm that :-
Stock FSB voltage , changes depending on CPU Type
65NM = 1.25v
45NM = 1.15v
Ace suggested that default (g)MCH of 1.25v made sence as it was aan intel standard , which ties up with my MM reading you saw in the link. On that basis i've also comnfirmed that there is an undervolt of ~0.20v on the overvolt. Confirming Anandtech's finding.
So apart from the mystery voltage from the middle pin next to the mch volts, which may be PCI-E volts, that's sorted. :)
Have updated original post ion link.
Team effort guys :) :up:Quote:
Thanks once again.
hey guys
whats the expected bandwidth for DDR1600 8-8-8-21 (7.5x400 q9450 x48t) ?
Everest gives me 8862mb/s and 8400mb/s
that is probably close for 400 fsb...as you increase the fsb, those numbers will go up since you are running the cpu system bus that much faster (L2 cache speed)
But its normal for the speed? because the comparative systems in sandra and everest show faster computers with only pc10000 ram and most DDR2's beat me...
just making sure lol
I want a bios for my X48T like this :) Gigabyte P45 Bios Facelift:shrug:
Quote:
Originally Posted by |Oc|REvO_HsArC
you should be able to tighten those timings up to 6-6-6-18 at 400 fsb.
You remember this post Celshock 14400 ddr3 died
Well Friday morning got the new kit Cellshock Blue Guess what , one of them's dead. Another 8~9 days :shrug:
What were the conditions they were running under when it occurred? Voltage, timings etc?Quote:
Originally Posted by Supershanks
As i say got them friday dropped them in, tried a quick 5 minute oc 1800 no real sress only +0.35, din't make it and I had a lot on so , just stuck them on stock , which believe it or not is jdec 1066 for the night. Tried in earnest with them yesterday. Tried a whole raft of settings based at 1st on my old set then with VDIMM: 1.80V = 1775mhz, VDIMM: 1.84V = 1810mhz, VDIMM: 1.90V = 1870mhz
from OC Report: Cellshock DDR3 1866 BLUE D9JNL
using the memset timings & voltage shown , also +0.10v extra
Was getting a post oiccasionally , but then it would rapidly lose stability with bsod or freaaze, Was freezing even at post screen on occasion.
I took a module out & it soared away , no worries. Swapped over to other module, won't post.
What a supreme bummer. That was a very unaggressive effort to have 1 fail so quickly.
Exactly , I was looking forward to getting a feel for them & then pushing on past rated 1866. So i'm baack to 1gb of 1333mhz :mad:Quote:
What a supreme bummer. That was a very unaggressive effort to have 1 fail so quickly
Seriously thinking of 2GB OCZ DDR3 PC3-16000 2000MHz Platinum EB Enhanced which I think are samsung, then flogging my blues if the ocz deliver, not sure :)
DARN!...that is the total pits with new memory dying on you like that!...it just goes to show you that even the best maker has bad luck with those micron modules.Quote:
Originally Posted by Supershanks
Yep it seems that these micron chips follow the tradition of DDR2, they die young:shrug:
Wouldn't mind if they'd had some chance to perform though :shocked:
it happened to me on Corsair DDR2 modules with micron chips...one stick failed within a couple of hours at a higher vdimm.
Hi guys
I have a little problem with my quadcore Q9450
See if you can tell from this printscreen, how good this cpu is: Its at 3.2ghz 1.134v
http://i30.tinypic.com/iwius8.jpg
yep, I have a bad core.
Three of them prime95 up to 18hours, no problem. The one with 64ºc Only primes 3 or 4 hours. If I feed him less, He fails quickly. If I feed him more, He fails quickly too.
Is there a way to Tune one core differently? GLT perhaps?
Your temps are overstated for a start by 10C see RealTemp
or Update Coretemp TJMax so it's right :)
GLt tuning may work see 20) GTLREF
luck:)
thanks :)
Click for more
http://www.images.clunk.org.uk/revie...v-1.3/v1.3.jpg
Supershanks and I were discussing where to readout MCH voltage...he thought that the readout for MCH could be measured from this Mosfet which is just on the backside of the NB bolt that holds the heatsink firmly over the chipset...this is between the I/O back panel and the NB heatsink
after setting all voltages to normal, i read with my DMM from the right hand post, which SS said was MCH voltage...i found it to be the "FSB" voltage, at least on my board...baseline or normal voltage was around 1.11v....and, it responded as i raised the FSB voltage add.
the middle post voltage actually moved up as i increased Vdimm...i had a slight variation from the Vdimm voltage reading over by the DIMM slots, but it definitely moved up and down as i adjusted the Vdimm...baseline or normal was around 1.53v.
new bios for x48t-dq6 after f5d ??
none that i heard about.
Got a new baby :shrug::D
I had to know you were going to when you were drooling all over that bios :ROTF:
Woooooo!...very nice that gigabyte gave you a sample to show off...hmmmm...coming up in the world!:up:Quote:
Originally Posted by Supershanks
We have to call him Sir now, or he doesn't acknowledge us peasants :lol:Quote:
Originally Posted by Ace-a-Rue
It does look a nice board though, I'm interested to see how that cooling does :)
:lol: yes my screen was in a right state :)Quote:
I had to know you were going to when you were drooling all over that bios
cheers Ace
Oh and Clunk :moon: :lol:
figuratively, we peasants, have to take a back seat!:DQuote:
Originally Posted by Clunk
he's misspelt sir he means cur :rofl:
i think he meant "Sir Mongrel":DQuote:
Originally Posted by Supershanks
Can anyone give me a quick heads up as to whether or not you can mount a Tuniq Tower on this board with that back-cooling on there? I've heard the story go both ways. Can someone clarify? Thanks a bunch.
-Collin-
i have a TRUE heatsink mounted on mine with a 120 x 38 MM fan which makes it like a Tuniq and there are no problems with any orientation...i would recommned removing the crazy cool back plate but loonym left the back plate on and he used store bought bolts to mount his TRUE heatsink....i believe he talks about it in this thread...start looking around page 4 and then go forward.
EDIT:..found the post and it was far away from page four!:D
http://www.xtremesystems.org/forums/...&postcount=286