• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Anyone OCing a Celeron D LGA775 2.66 on a V915P???

D

dripgoss

Senior member
OK, I decided to experiment a bit by building a budget PCIE rig. Among other things, I bought a Celeron D 330J 2.66 LGA775 E0 stepping proc and a Chaintech V915P board. I've googled for days on these components but since this is a pretty new platform combo, I'm coming up dry on suggestions for overcoming my hurdles in OCing this thing so I need your help...

The problem is, when I try to OC by just increasing FSB in the BIOS, I can only get it to 140 and be able to boot into Windows (2.8Ghz). If I go past 140 and save changes in the BIOS, the board just simply resets back to 133x20 (2.66) and reboots - yet it keeps all my other BIOS tweaks like voltages, RAM timings, etc. Default CPU voltage is like 1.28, but I've tried as high as 1.325. I'm affraid to go much higher, but even overvolted the CPU reports 45°C which is supposed to be great for Presscott cores...

I've read posts about "memory holes" on the 915P/Deleron platforms, but I can't find any hard fast suggestions. I've tried FSB's from 140 to 200 and the only 2 RAM ratios my mobo supports (400 and 333) to no avail. Also tried locking the PCI freq. to 33 and PCIE to 100 which are the only locking values.

I don't think it's temp related as I have a 35°C ambient and the proc stays between 39°C and 45°C consistantly (via BIOS HW monitor, I don't have a diode installed yet). I don't think it's RAM related, as I've relaxed all my timings and tried even bumping vdimm +0.02 for both sets of RAM I've tried (see below for specs). My PSU is supposedly decent although I haven't found a supported Windows volage monitor and I'm not sure how much juice these Delerons need.

Here are my specs:

Celeron D 330J 2.66 E0 Stepping w/ Stock HSF & AS5
Chaintech V915P LGA 775
(2x) 512 Corsair ValueSelect DDR400, Matched, DC config (SPD is 2.5-3-3-6, tried 3-4-4-10)
(2x) 512 Kingmax Hardcore Series DDR500, Unmatched, DC config (SPD is 3-4-4-8, tried 3-5-5-10/12)
Thermaltake Pure SilentPower 420W PSU
MSI 6600GT 128MB PCIE
Hitachi 80GB SATA HDD
MSI CDRW/DVD SATA Combo
Skyhawk Al Case w/ plenty of fannage

So is there anything I might be missing? Could it be the new NV thing or execute disable bit crap? Is it my RAM or PSU? Any suggestions? I've read some ppl getting 200FSB outta their 478 D0 Delerons, but is the new LGA775 Deleron's just borked? 🙂

Thanks for any input and Happy New Year!!
 
OK, just an update in case anyone else is running this combo and decides to search...

I tried just running a single 512 stick in case it was a DC issue. And now, the system will boot at 2.8Ghz 140 FSB into Windows but when I shut down and reboot SOMETIMES it goes back to defaults (133FSB). I haven't loaded Prime95 yet, but I can play 30 minutes worth of Doom 3 @ 2.8 with no problems, rebooting or heat issues. In fact no matter what I throw at this thing, the proc stays under 47°C. I'm thinking this could be a BIOS bug or something due to the fact that it just sporadically resets whenever the heck it feels like it.

I tried taking it to 1.325v which I'm afraid is close to the edge for this proc since they use less voltage than other Prescotts or the Celeron D 478 counterparts.

Is it just that the newer Celerons can't handle higher than 533 system FSB? Does anyone know actually what voltage these will take?
 
Have you read any overclocking faq's? You may need to disable spread spectrum, and set the memory below 400 to overclock it. I would also use "spd" to start for memory timings, and see what the board selects for your memory. Some boards use "ratios" for memory settings; and some use 266,333, 400, etc. You should be able to run the celeron at 1.4 volts, especially if the board is slightly undervolted. When you set and save your settings, see what the bios shows for cpu voltage in the "pc health" or similar section.
 
Thanks o1die for the tips, but I didn't mention I have disabled spread spectrum, set RAM to the only divider the board gives me 333 (5:4), I'm definitely keeping the RAM at or below DDR400, and I included the SPD/Tried timing's of the two sets of RAM above because that's what the mobo detected vs. what I relaxed to. Besides, even though I'm not an uber-l33t-ninja OCer, I've been OCing AMD's for years and have read countless OCing FAQ's over the years... 🙂

As for PC Health, when the mobo does keep my mild 2.8 OC on boot, PC Health shows everything the way I set it. However, after I reboot a few times, the BIOS resets back to default FSB but will leave all of my other tweaks untouched. So I'm thinking one ro more of the following are possibilities:

1.) These E0 Delerons either don't like +533FSB (which I've read should NOT be the case)
2.) I'm not getting enough voltage to the CPU (possible but again, I don't wanna fry the chip and +0.1v is scary on air for Intel)
3.) The MB doesn't like either set or RAM (doubtful, but possible I'e tried some pretty loose timings and even tried +0.2v)
4.) I'm hitting one of those "FSB walls"

Things I'm pretty sure are not the culprit:

1.) My PSU isn't good enough (everything is rock at 140 FSB but I can't even hit 141 without the BIOS resetting)
2.) Things are over heating (unless my PC Health/HW Mon in the BIOS is totally borked, CPU temps are below 47°C no matter what I do and ambient is around 32°C)
3.) PCI/PCIE locks aren't working (pretty sure this isn't the case because I haven't corrupted my HDD or fried my GPU yet)

I know I seem to be having a one-man conversation here, but I'm just using this thread to jot down my thoughts on this and if anyone sees any blaringly obvious factors I missed, thanks for posting your thoughts...
 
It really sounds like a simple case of a crappy BIOS that needs to be updated. You are at the mercy of Chaintech's BIOS engineers. 🙁
 
Thanks Mega..

Yeah, Chaintech was pretty decent about getting updates out on my VNF3-250, but this board has been out for a few months and not a peep from them. My manderine is a bit rusty (OK, non-existent!) but I even tried plucking around their Chinese site to see if they had a BIOS that just never made it to the English downloads section...

The thing that kills me is that the reviews I read were talking about how well this board OC'ed with D0 presscott's (albeit P4's) and I read where the Deleron 2.66's OCed well also. I figured the lower heat of the E0 and smaller cache would help me with the temps and so far I was right about that at least... If I can figure out what the hell the deal is, this could be a nice little budget PCIE system - if not oh well. I'll chuck it and go NV4/939 anyway!! At any rate I like the challenge...

I'm wondering if the new halt technology on the E0's isn't causing some issues. I'm trying to read up on what this does anyway...
 
OK, I took it up to 1.385v and now that my AS5 has had a chance to settle in, my chip is at 39°C to 45°C even overvolted! I still can't hit any of the FSB's above 140 though. I've read some others hitting issues like this with other 915 boards but with C0 and D0 stepping non-LGA Delerons so maybe it's either something with this chipset or something with the Celeron D's?

Then again, I turn around and read some guys are hitting 173-200 FSB with these things on SIS platforms so WTF?!?! I know my 2.66 won't run at 4Ghz, but it would be nice to get some kinda OC since my temps are so low and this thing obviously doesn't mind the voltage...
 
OK, for anyone else who might be in the same boat, this little piece on THW explains a possible issue regarding OCing an E0 chip:

The processor is protected against overheating by Thermal Monitoring 2, which is activated by the so-called PROCHOT signal. This causes the CPU to reduce the clock if that is needed to cool things down, while the core voltage is simultaneously lowered to 1.2 V. On top of that, the CPU can fall back on PROCHOT whenever it runs idle. This happens via Enhanced Halt Mode C1E as soon as the operating system's CPU demands drop to low values.
Click to expand...
(Full article here.)

I wonder if there is a way to disable the TM1/TM2 on these or if that's even recommended or necessary..

I'll do more searching around tonight. and the saga continues...
 
OK, I went up to 1.425 on the volts and I still can't get this thing to keep an OC. Now, even if I try 140FSB and 5:4 mem ratio, relaxed RAM, put a cooler on the NB, it still doesn't want to hold the settings on boot. On the rare occassion it DOES boot into Windows @ 2.8, it loses it on next shutdown.

BTW - I found a very curious thing in the V915P BIOS in the "CPU Feature" menu. It's the otion to disable NX virus protection or something. Now, I don't know much about this and tried to disable it (to no immediate effect), but it did seem to make my system boot faster. Also in this menu was a grayed out item that said "T1 Thermal"...

I'm still thinking this thing just has a buggy BIOS. I've sent Chaintech a few emails, but hopefully they're working on a new BIOS rev for this mobo. Otherwise, this has been a very stable and problem free install and I'm gaming on it with very comfortable temps for a modern Intel system using stock cooling...
 
Hory Clap! It looks like I got the right proc, wrong mobo...

This Japanese Guy is hitting 4.0Ghz & 800Mhz FSB with my proc (albeit on a 925x Abit AA8-DURAMax ). Of course the translation is a little rough, but with the pix and what I could make out, this could still be a worthwhile experiment with the right board. I mean, the temps and voltage tolerances are so right, but 915P is seeming so wrong...

As said earlier, I like the challenge and don't mind uncharted territory, but part of me can't wait until NF4/939/PCIe has proven itself a bit before I jump....
 
More notes to self:

I crossed my fingers and took the CPU voltage up to 1.5 knowing full well that I might get a meltdown and you know what happened? Nothing! My temps didn't even fluctuate from the ~41°C idle I've been getting at 1.285. The only thing that I noticed was my 12v rail drooped from stable 12.5v to 11.8v at times. I even ran Prime95 for a little while and played a few rounds of Halo! I still couldn't get any FSB's above 140... I left everything else stock BTW - with relaxed RAM, SPD it didn't matter...

I did this because I wanted to rule out the possibility of not enough voltage holding me back. I think 1.5v on a chip that defaults to 1.285 should be plenty. So, I guess I need to wait for a BIOS update, although since this isn't a high priority platform for the hardware industry, I might get screwed... Oh well...
 
Finally got a chance to swipe the diode and LCD from my wifes case (shhh!) and do a hardware test on the temps. At the HSF base next to CPU socket, it shows 39°C idle! I'm amazed so far at how unhot this processor really is compared to all the "OMFG - PRESSCOTT EASY BAKE OVEN!!!" horror stories I've read. I know that has a lot to do with 256k cache, but who cares?

I even tried checking around the Chaintech Korea site (Google translate actually works, somewhat...) and can't find an updated BIOS yet. Kinda surprising since CT is usually pretty decent about BIOS and driver files. But they need to fire their US webmaster because drivers take forever to make it from the asia sites to the english one!! heh...

I'm gonna give this sucker a rest for a few days and I'll go back to OCing my VNF3-250/3000+ to clear my head...
 
Originally posted by: dripgoss
Hory Clap! It looks like I got the right proc, wrong mobo...

This Japanese Guy is hitting 4.0Ghz & 800Mhz FSB with my proc (albeit on a 925x Abit AA8-DURAMax ).
Click to expand...

Actually I think you got wrong mobo AND wrong CPU. For the motherboard, I've seen plenty of times where the same CPU/RAM/etc on one motherboard OC higher than another motherboard. As for the CPU, my reasoning goes like this...

1) Maximum that manufacturer is getting from process is 3.8GHz, meaning that is attainable with a slower chip.
2) Motherboards seem to like running at "default" speeds. For instance I have a Deleron 320 2.4GHz that runs fine at 3.6GHz on my Asus P4P800 Deluxe with BIOS set to 200MHz. If I set BIOS to 199MHz, won't POST. Either it is a FSB "hole" or the memory multiplier gets set way out of whack (known on the P4P8X at the 199/200MHz juncture). I'm using Kingston HyperX PC4000, but still possible for motherboard to set multiplier to run RAM over spec.
3) Using "default" 200MHz FSB you'd have to run at 4GHz with your chip. This is above the maximum that Intel sells, making it a more unlikely overclock (though still possible). The only other Deleron I've had besides the 320 is a 330. It would NOT run stable at 4GHz. It would run, but very unstable. This was also on my Asus P4P800 Deluxe with Kingston HyperX PC4000 and a Thermalright XP120 with a Panaflo 120x38mm fan. Since my motherboard would not POST from 199MHz down to around 160-170MHz, the possibilities of the 330 (2.66GHz like yours) hitting 3.6-3.8GHz are low because of the motherboard and not necessarily the CPU. However, this makes your choice of CPU "wrong" in that you should have gotten one with a lower multiplier.
 
Yeah, you're right I'm sure. This was PCIe stepping stone system at best - and an experiment that didn't work at worst. It sure fit the budget at the moment and it games fairly well, which was all I was really interested in anyway. So I'll proly just FS/FT it next month and go with an NF4 solution.
 
I forgot to post up here, but for anyone interested, I was able to get to 3.32Ghz stable on this 2.66 using a version of ClockGen made for 915 chipset. In the BIOS, I locked down the RAM ratio to 333 (5:4) and the PCI to 33Mhz, left default 133 FSB, bumped up the vcore to 1.45v and the NB and vdimm to +.2v, disabled Spread Spectrum and used SPD timings (lose timings didn't raise stability). However, I read one of the AT reviews of the 915/925 chipsets from last year talking about the PCIE lock being non-functional. And they were right! My temps are great, everything is stable and all indications are that this things hould go higher - HOWEVER the OC craps out when my PCIE bus reaches around 120Mhz (Mines at 124 when proc is at 3.32Ghz) which is what they found as well. It appears some of the other boards on this chipset can get around this limitation, but obviously not my V915P - or at least not yet. If anyone knows how to lock the PCIE, do share because I think this thing would go higher without that problem.

After tweaking my 6600GT to 575/1100, my 3DMark 2001 SE went from 12000 to 15024 and I'm seeing noticable improvements in game. Temps aren't bad - OC'ed, CPU is 42°C/56°C ambient around 35°C. GPU gets toasty though - up to 71°C during select games, but most of the time around mid 60's at load so I won't work too much...

Just thought I'd update...
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top