- Joined
- May 26, 2004
- Location
- San Diego, CA
I keep getting excuses from him, its really p issing me off.
-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
Official Presler/Cedar Mill Overclocking & Issues Thread
- Thread starter Sentential
- Start date
Clearly OT, but he's always been slow about getting things out. I warned proth before he ordered his 
In addition to there not being enough hours in the day, he's had some situations lately that have made it even harder for him to them done. I'd just keep PMing him and eventually you'll each get them.
In addition to there not being enough hours in the day, he's had some situations lately that have made it even harder for him to them done. I'd just keep PMing him and eventually you'll each get them.- Joined
- Feb 4, 2003
- Location
- SF Bay Area
2 weeks would be extremely expedient by chilly's standards.proth said:
- Joined
- Jul 21, 2002
- Location
- Branford, CT USA
I'd have to research the type of processor your using to fully appreciate it's "behavior" ("Bad processor did I catch you Thermal Throttling yourself again! Go to your Box!") I'm also unfamiliar with motherboard for 805s. Aren't these usually Dual Socket boards? Tyan etc. From what I know of Tyan they probably have some solid power stages on their boards. I don't thjink that's the issue I do think it may be related to your Antec 430W PSU, which may be underpowered for the application, especially if the modo is dual socket.goog` said:
As far as I know EPS12V (perhaps the ATX as well) specs are actually under revision, their requirement (at least the EPS12V) has been unoffcially dropped. What manufacturer would risk a potentially substantial loss in Sales if they cannot claim they meet Server System Infrastructure and Form Factor guidelines even if these guidelines are questionable? As an anology Mobil introduces Mobil One which was lubrocant based on graphite. The result was the longest lasting, most effective engine lubricant on the market, but they can't call it Motor Oil.
Comply or die. The exception to that rule would be a PSU whch is itself exceptional, this is where PCPower&Cooling come into play. Look at any store which sells their products, most of their models do not meet certain EPS12V guidelines, yet out perform anything in their class.
Monopolies such as Intel and Microsoft dominate the industry and while any engineer not smoking crack knows, Intel's desktop CPU design architecture had come to a fork in the road. One path followed PIII (shorter pipeline lengths more IPCs, low power consumption (heat)) and down the other path Pentium (with it'S NetBurst nightmare) they chose Pentium and made some decent money exploiting the laypersons unfamiliartity with CPU design. Speed is what people wanted and speed is what people got, so what if they never got out of first gear, it sounded fast and in some respects did run fast (internally). There were some innovative CPU design features coming out of Santa Clara such as Voltage Regulator Down and On Demand which would have been cutting edge on their own merits but they were invented only as ad hoc fixes for a processor which ran as hot as a nuclear reactor. Three Mile Island had nothing on Intel and Dual Core has only made matters worse.
Insofar as Socket-478 Prescott, many will disagree with this staement but beyond the heat issues which are par for the course for Intel, tey were excellent overclocker's and chip yeild was excellent. All these processors scaled well what held 99% of the CPU's back were ineffective motherboard power circuits. Thi is made clear in my article Prescott Effects On Motherboard Power Circuits, also [H]'s article Presoctt Tweaks Dangerous and finally IAmNotAGeek's Prescott Info unfortunately that article was supported with refernces to Inel's Design Guidelines such as the 478-VRD White Paper (Sec 2.4). [H] as large as have access to ore-production models and this is where fundamental design flaws are most obvious. they actually melted plstic when overclocking their Socket-478 (the worlds first 90nm desktop) whentesting the chip. Of course thy had layed the board flat with no air-flow beneath, nonetheless by the temps recorded using the Abit AI7 which I purchased because of it's unique placement of monitoring devices were exceeding 158F (70C)esigation makes these 65nm CPUs the Fetus in Fetu of the modern processor world. These are failed Presler Dual Cores. I'm struggling with the fact my Pentium-D 930 doesn't have hyperthreading enbaled which frustrates me to no end, I couldn't deal knowing I'm carying around a dead core. I'd wait for Conroie which will be here soon enough. Trust me save up for that CPU because in DC with a motherboard to support it your going need every penny! If you want a decent LGA 775 chip get a Prescott 630. Mine easily does 4.0Ghz without much if any Vcore increae at times depending on H20 or aircooling.
Attachments
Last edited:
- Joined
- Jul 21, 2002
- Location
- Branford, CT USA
As far as Chilly he repaired two Prometeias sent to me from Chip-Con whom I gave to my buddy JNAV. In both instances the dam cpaillary tubes had leaks. He did a great job
Anyway it took a bit of time but he kept all informed. I think many of the older units are now due for a tune-up etc and people are sending them out en mass.
On the subject of Phase Change I wonder how OCZ's units re doing?
Anyway it took a bit of time but he kept all informed. I think many of the older units are now due for a tune-up etc and people are sending them out en mass.
On the subject of Phase Change I wonder how OCZ's units re doing?
Ever seen the movie "The Money Pit"? That's the 2 weeks he meant
Chilly does great work, but so do others.OCZ Cryo thingie =
They built it (at least some number of "testers") and supposedly can't figure out how to ship them without the compressors ripping off the base when they're handled roughly during shipping. Comp comes lose, copper tubes crack and they arrive completely degassed. Comps are heavy of course, but I can't see why putting some wide fender washers between the bolts and the bottom of the case on the outside wouldn't be enough to spread the load enough to keep them from ripping through?Anyway, supposedly they'll be available around the time Conroe is released, but a lot of people are already calling it "vaporware". I'll believe it when I see it.
- Joined
- Jul 21, 2002
- Location
- Branford, CT USA
Thanks reminds me of what happened with my Prommie's. Morten was kind enough to give the me the review sample after it made the rounds. It arrived damaged, insurance paid for brand new unit and it worked for about a week after the gas leaked out. One thing was evident, all they needed to do is use a small pallet. These shipping pallets along with Fragile stickers are the reason so many VapoChills maike it their detinations without damage. At least I bvelaive they were shipping them that way. I suggested to Chip-Con with their market share growing in the USA they employ a small shipping pallet mband strapped to every unit. But I'm just a little ant in the insect world.
Even if they fully crated them, the problem is, there is space inside the case itself, so the comp itself really needs to be held. If they could find a way to bolt the comp through the case into a wood base, that would work much better, but I am guessing they don't want people to have to open the case at all when it arrives. They need to spread the load on the bolts (wide washers) or maybe use a thicker sheet metal for the base. I heard they had tried shipping screws to hold the comp inside the case with moderate success, but extra screws or holes in the case would probably be pretty unsightly. I don't know...they'll need to figure something out themselves, I am not on the payroll
- Joined
- Jul 21, 2002
- Location
- Branford, CT USA
Well I agree, i wasn't trying to rsolve the problem either. It's almost always abou cost. Since it would take so many man-hours to drill, bolt, package the units, even supply a small wrench to remove the fastening bolts and print instructions showing how to remove the shipping hardware before operation then you've just added $50 to the price of a unit wagering it's success on keeping it's cost under $500 and then $400. Just having special pallets constructed would be a hassle, if not prohibtively costly.
Last edited:
I hear ya! On a good note, Victor is running his Conroe (stock) with a passive heatpipe sink (not sure which) and it doesn't go over 35C on load, 25C idle in 22C ambient. He wasn't trying to test passive cooling. There was a fan on the sink, but sometime after the fact, he realized he never plugged it in. I am sure they'll be much warmer when OCed, but had that been a Prescott, it would've thermally shut down 5 secs. after the first bench started, LOL.
- Joined
- Jul 21, 2002
- Location
- Branford, CT USA
Funny you say that i was just going to say we're forced by necessity to focus a large part of our Overclocking energy into thermodynamics. Were approaching a problem from an ad hoc perspective.
I do believe many of us are excited about "playing" with these toys and many would actually miss their H20 systems and phase-change units if we were back in the days when passive HS were all the rage (boring). So obviosuly we derive a great pleasure form this. Have no fear we'll still be seeking ways to cool our GPUs for some time and I wonder what the effects will be when we really begin delving into cooling storage devives which run exceedingly hot and have so many moving parts? There's plenty to "cool".
Conroe is the best decision Intel has made since the PIII. Temps will soon drop and where H20 system only placated the the two-headed beast known as Presler, the benefits plain ole H20 cooling will have on Conroe would be anlogous to the effects phase-change now have on Presler. Sure speeds will be lower as will many things (temps) yet the amount of work, the IPCs that truly matter will increase. Everything else is like masking your problems with big fuzzy dice, if you stare at them you forget your driving an NPR Car-Talk case even Click and Clack couldn't help you with.
I do believe many of us are excited about "playing" with these toys and many would actually miss their H20 systems and phase-change units if we were back in the days when passive HS were all the rage (boring). So obviosuly we derive a great pleasure form this. Have no fear we'll still be seeking ways to cool our GPUs for some time and I wonder what the effects will be when we really begin delving into cooling storage devives which run exceedingly hot and have so many moving parts? There's plenty to "cool".
Conroe is the best decision Intel has made since the PIII. Temps will soon drop and where H20 system only placated the the two-headed beast known as Presler, the benefits plain ole H20 cooling will have on Conroe would be anlogous to the effects phase-change now have on Presler. Sure speeds will be lower as will many things (temps) yet the amount of work, the IPCs that truly matter will increase. Everything else is like masking your problems with big fuzzy dice, if you stare at them you forget your driving an NPR Car-Talk case even Click and Clack couldn't help you with.
Last edited:
Ad hoc, but you gotta work with what you're given. Intel chose to bestow furnaces on us for the last 4 years Conroe will be MUCH cooler to start, but don't think they won't get warm when pushed to the limit. Of course we don't know how high they will OC yet, but the main factors will be FSB and Vc. If it gets anywhere >4GHz or requires 1.55Vc+ to do it, it could get darn hot. Considering the FSB they will be pushing for a 4GHz clock, the 2.66GHz would be running 400FSB which is doable, but a 2.40GHz would be running a pretty hefty 444FSB. The NB might overheat long before the proc, LOL. If they do clock well, mobos might be the OC limiting factor and NB cooling might become just as important as the CPU.
The 4 instructions are very nice (it's actually 5 in, combined to 4 out sometimes!), but you really should look at the rest of the core...there's a lot of cool stuff in it.
Conroe will be MUCH cooler to start, but don't think they won't get warm when pushed to the limit. Of course we don't know how high they will OC yet, but the main factors will be FSB and Vc. If it gets anywhere >4GHz or requires 1.55Vc+ to do it, it could get darn hot. Considering the FSB they will be pushing for a 4GHz clock, the 2.66GHz would be running 400FSB which is doable, but a 2.40GHz would be running a pretty hefty 444FSB. The NB might overheat long before the proc, LOL. If they do clock well, mobos might be the OC limiting factor and NB cooling might become just as important as the CPU.The 4 instructions are very nice (it's actually 5 in, combined to 4 out sometimes!), but you really should look at the rest of the core...there's a lot of cool stuff in it.
- Joined
- Jul 21, 2002
- Location
- Branford, CT USA
Oh I'm psyched it's got some real potential! I think your spot on the issue will certainly be NB related and lets be honest until the 975X (which i don't think people realize has some hidden potential) the 875 was it. If it's true the 975x will be the first chipset supoorting Conroe then 775 will be the first package? I did find the 975X to improve the graphics through-put. In fact I ran a 6800GT just to be sure.
I think cooling the NB will be key with Conroe as well as cooling the CPU what i don't like is that I understand Conroe will be released on a 775 package. I'm finding these "sockets" do seem to have a finite life. Some boards show Socket pin damage after about ten installations and that's being very careful. You have to install the CPU with the board laying flat, if the CPU slips or shifts while mounted in a case the CPU "drops" catching pins with it and bending them out of place.
I think cooling the NB will be key with Conroe as well as cooling the CPU what i don't like is that I understand Conroe will be released on a 775 package. I'm finding these "sockets" do seem to have a finite life. Some boards show Socket pin damage after about ten installations and that's being very careful. You have to install the CPU with the board laying flat, if the CPU slips or shifts while mounted in a case the CPU "drops" catching pins with it and bending them out of place.
Yes, the 975 chipset supports Conroe, but current boards with it don't because of incompatable power circuits. Board revisions are coming, along with some new purpose-built (Conroe) 975 boards. New 965 boards will be also be released near Conroe and I've yet to see the full specs on the chipset, but the 975 chipset will likely be the premium chipset until the 985 is released near the end of the year.
And yes, Conroe is 775 socket. Talk about funny you should mention that, not even yesterday I was talking to someone about the socket and recalling that when they were released people were talking about them only being good for limited installs. IIRC, people were saying 10-20x before the socket/board would need to be replaced. I found it funny as I mentioned that I easily have 5x that amount of installs on mine and routinely torque evaps and DI containers enough to warrant a 1/2" torque wrench OK, not that much, but enough to bend the board and I've had no problems with pins yet. I do see people messing them up just installing procs, packing the sockets full of DEG or dropping something on the pins, but short of that, mine has held up very well and I am not particularly easy on it.
As for the FSB, we'll need to see what the new boards are like. I am hoping we'll see a bump in max FSB from 450->500 and that might require some improved NB cooling unless the chipsets receive some refinements and/or run reduced voltage too. Many of the Conroe-specific boards I've seen so far are 8-phase power and have switched to digital PWM/PLLs. Clearly the emphasis with the new cores is low, clean and tightly regulated power. That will reduce heat everywhere, especially in the proc. A digital clock generator in place of a standard PLL chip is a whole lot more stable too and I've seen 400-500MHz increases in clocks on the same proc between a board with standard PLL chip and digital clock. More to the point, overclocks are not only higher, but much more stable. Keep in mind that everything on the board that has anything to do with bus speeds all originate with the PLL chip and the closer it gets to "max", the more unstable everything becomes. Digital does not have that issue and FSB will be limited by what other components can stand, most likely the NB.
And yes, Conroe is 775 socket. Talk about funny you should mention that, not even yesterday I was talking to someone about the socket and recalling that when they were released people were talking about them only being good for limited installs. IIRC, people were saying 10-20x before the socket/board would need to be replaced. I found it funny as I mentioned that I easily have 5x that amount of installs on mine and routinely torque evaps and DI containers enough to warrant a 1/2" torque wrench
OK, not that much, but enough to bend the board and I've had no problems with pins yet. I do see people messing them up just installing procs, packing the sockets full of DEG or dropping something on the pins, but short of that, mine has held up very well and I am not particularly easy on it.As for the FSB, we'll need to see what the new boards are like. I am hoping we'll see a bump in max FSB from 450->500 and that might require some improved NB cooling unless the chipsets receive some refinements and/or run reduced voltage too. Many of the Conroe-specific boards I've seen so far are 8-phase power and have switched to digital PWM/PLLs. Clearly the emphasis with the new cores is low, clean and tightly regulated power. That will reduce heat everywhere, especially in the proc. A digital clock generator in place of a standard PLL chip is a whole lot more stable too and I've seen 400-500MHz increases in clocks on the same proc between a board with standard PLL chip and digital clock. More to the point, overclocks are not only higher, but much more stable. Keep in mind that everything on the board that has anything to do with bus speeds all originate with the PLL chip and the closer it gets to "max", the more unstable everything becomes. Digital does not have that issue and FSB will be limited by what other components can stand, most likely the NB.
- Joined
- Nov 16, 2005
- Location
- upp north, europe
Ok, now my pressler is realy starting to pi** me off...
System is: P D 950es, P5WD2-E Prem, Corsair Pc8000ul (D9) Tagan Tg580w,watercooled..
Problem: sudden reboots. Can suddenly reboot.. idle or under load... when it reboots the next boot often end up in a bsd... but if i go into bios and lay off some v-core and mhz it might boot ok.. but problem also apper at stoc speed... fsb seem to not be the thing, nither mem-speed.. ore v-core.. starting to think my chip is totaly F...ed...
reinstalled windows ocd to [email protected] bios... ran sp2005 + rthdribl to give 100% load at cpu and gfx.. psu seem to be ok.... rail drop to 11,5v idle is app 12,5(mulitmeter)
it ran for 14h and when i shut down the load apps and started surfing the web. i got a reboot...
have not tryed memtest, but dont expect my mem to be the problem, have tryed hi and low speeds and timings, no difference
have tryed sh**tloads of things... waiting for a p5wd2 prem from rma and will test it also but im starting to think about a new cpu...
Help...
System is: P D 950es, P5WD2-E Prem, Corsair Pc8000ul (D9) Tagan Tg580w,watercooled..
Problem: sudden reboots. Can suddenly reboot.. idle or under load... when it reboots the next boot often end up in a bsd... but if i go into bios and lay off some v-core and mhz it might boot ok.. but problem also apper at stoc speed... fsb seem to not be the thing, nither mem-speed.. ore v-core.. starting to think my chip is totaly F...ed...
reinstalled windows ocd to [email protected] bios... ran sp2005 + rthdribl to give 100% load at cpu and gfx.. psu seem to be ok.... rail drop to 11,5v idle is app 12,5(mulitmeter)
it ran for 14h and when i shut down the load apps and started surfing the web. i got a reboot...
have not tryed memtest, but dont expect my mem to be the problem, have tryed hi and low speeds and timings, no difference
have tryed sh**tloads of things... waiting for a p5wd2 prem from rma and will test it also but im starting to think about a new cpu...
Help...
- Joined
- Jul 21, 2002
- Location
- Branford, CT USA
Tagan's TG580-U22 follows the EPS12V Power Supply Guide 2.1 (split plane), 12V1 and 12V2 each operate at 20A. This does not imply 40A are avilable to the PSU.
Your Presler is a hungry beast, and just as an experiment if you can borrow an older PSU pre ATX 2.01 / EPS 2.1 guide revisions with at least 33A on the 12V Rail this will at least answer the question if in fact th PSU is the culprit.
Tagan builds an excellent PSU every PowerStream uses a Tagan licensed design and it's the OCZ PowerStream 520W which has been helping many Presler owners, although just barely with a single 34A 12V rail.
Tagen is releasing thei TurbioJets 900W and 1100W PSU's , I imagine these have ma greater flow rate (amperes) in the 12V current department.
Another problem may be total drain on the PSU; if your watercooling the PSU what else are you watercooling? Between case-fans, radiator-fans and pump/s this can drain 12V rails and your Tagan may be fine with a minimal number of compoments, but the added H20 hardware while necessary, may be stressing the PSU in combination with Presler.
The other thoughtbn that came to mind was BSOD's are often memory related and the randomness of your re-bott may not be so. I would try running the PC2-8000 on SPD (or AUTO depending on the BIOS). These are not modules that do well under manual settings running 1000MHz regardless of timings because at this speed they run better at the manufactuers pre-progammed (SPD) timings. They tested or stability at 1000MHz on SPD. The same IC's in your PC2-800 are most likely those found in all PC2-6400 as well, these simply binned highest. This is their operational ceiling at loose timings.
Try SPD first, CMD2 and if the DDR voltage is above 2.15V try putting it on AUTO as well.
Your Presler is a hungry beast, and just as an experiment if you can borrow an older PSU pre ATX 2.01 / EPS 2.1 guide revisions with at least 33A on the 12V Rail this will at least answer the question if in fact th PSU is the culprit.
Tagan builds an excellent PSU every PowerStream uses a Tagan licensed design and it's the OCZ PowerStream 520W which has been helping many Presler owners, although just barely with a single 34A 12V rail.
Tagen is releasing thei TurbioJets 900W and 1100W PSU's , I imagine these have ma greater flow rate (amperes) in the 12V current department.
Another problem may be total drain on the PSU; if your watercooling the PSU what else are you watercooling? Between case-fans, radiator-fans and pump/s this can drain 12V rails and your Tagan may be fine with a minimal number of compoments, but the added H20 hardware while necessary, may be stressing the PSU in combination with Presler.
The other thoughtbn that came to mind was BSOD's are often memory related and the randomness of your re-bott may not be so. I would try running the PC2-8000 on SPD (or AUTO depending on the BIOS). These are not modules that do well under manual settings running 1000MHz regardless of timings because at this speed they run better at the manufactuers pre-progammed (SPD) timings. They tested or stability at 1000MHz on SPD. The same IC's in your PC2-800 are most likely those found in all PC2-6400 as well, these simply binned highest. This is their operational ceiling at loose timings.
Try SPD first, CMD2 and if the DDR voltage is above 2.15V try putting it on AUTO as well.
- Joined
- Nov 16, 2005
- Location
- upp north, europe
Liquid3D said:
Well the tagan is "split" atx2.01 20A+20A 12v or "Combine" Atx1.3 35A 12v i run combine all the time..
the load on the tagan is 950d+mb+7800gtx+5x120mm fan+2xhdd+pci sound and a floppy, water cooling is 220v should not be to much, at least not running stock speed. those reboots are weird.
Just found a OCZ PowerStream 520W at my local forum and will be getting it..
will also test the memory more... relay hope it is ok... it is killer mem..
tanks for the tips, anyone else got something to contribute with?
Last edited:
- Joined
- Jul 21, 2002
- Location
- Branford, CT USA
I would hold off on the PowerStream because PCPower&Cooling is supposedly coming out with a very high ampere 12V rail PSU which is below 800W, ergo the price is affordable.
I'm currently testing the XG Duro 900 12Vx4 I'm on the SLI stage and heading into Presler testing on the P5WD2-P.
I'm also testing the PCPower&Cooling 1kw which is pricey obviously, and requesting the Tagan Turbojets 90W and 110W the 900W should be affordable to most Enthusiasts. If your spending $100 I'd wait and spend $50 more for a PSU which will last for sometime. More cvurrent more control, more current ,ore peripherials such as H20 cooling. It's better to have too much then too little.
I'd be willing to bet my now infamous Solder Dropping collection (with three faces of Jesus) there's going to be a Powersteam 950 and 1200W (since they like to be on top but can only stretch the Tagan circuitry so far) in the next few weeks based off the Tagan 900W/1100W topology, except with OCZ added features external pots etc.
PS: "combine" mode isn't the same as a dedicated mode as far as I know, this i why the Powerstream 520W which would technically offer less current at 34A are working so well.
Have you Mem-tested your PC2-8000 sticks? Have you tried dropping voltage to Auto and timing's to SPD?
I'm currently testing the XG Duro 900 12Vx4 I'm on the SLI stage and heading into Presler testing on the P5WD2-P.
I'm also testing the PCPower&Cooling 1kw which is pricey obviously, and requesting the Tagan Turbojets 90W and 110W the 900W should be affordable to most Enthusiasts. If your spending $100 I'd wait and spend $50 more for a PSU which will last for sometime. More cvurrent more control, more current ,ore peripherials such as H20 cooling. It's better to have too much then too little.
I'd be willing to bet my now infamous Solder Dropping collection (with three faces of Jesus) there's going to be a Powersteam 950 and 1200W (since they like to be on top but can only stretch the Tagan circuitry so far) in the next few weeks based off the Tagan 900W/1100W topology, except with OCZ added features external pots etc.
PS: "combine" mode isn't the same as a dedicated mode as far as I know, this i why the Powerstream 520W which would technically offer less current at 34A are working so well.
Have you Mem-tested your PC2-8000 sticks? Have you tried dropping voltage to Auto and timing's to SPD?
- Joined
- Nov 16, 2005
- Location
- upp north, europe
Ran memtest for 8 hours @ DDR1000 cl5-4-4-9@2,2v (what corsair say is spd, but auto give 5-6-6-18)
Mem is clearly not the problem.
Must run low cpu speed to not get errors, if i oc to high(4-4,5ghz 1,52v) it will fail within minutes. but runs ok at lower speed (approx 3-3,8ghz)
So i got realy frustrated and ripped the system totalt appart... and i might have found the problem... at some point i have made a small dent/mark in the ihs.. and it might have caused bad contact over one core or so.. but temps have alvays been fine (sub 50 allways.)
So i have sanded down the dent and will now start testing and provoking the system... feel totaly like a noob now...
OK... No more reason to feel like a noob... the problem is back... Now i started with f@h x2 @ 4.6ghz for 2 days... and now suddenly the rebooting is back....
Mem is clearly not the problem.
Must run low cpu speed to not get errors, if i oc to high(4-4,5ghz 1,52v) it will fail within minutes. but runs ok at lower speed (approx 3-3,8ghz)
So i got realy frustrated and ripped the system totalt appart... and i might have found the problem... at some point i have made a small dent/mark in the ihs.. and it might have caused bad contact over one core or so.. but temps have alvays been fine (sub 50 allways.)
So i have sanded down the dent and will now start testing and provoking the system... feel totaly like a noob now...
OK... No more reason to feel like a noob... the problem is back... Now i started with f@h x2 @ 4.6ghz for 2 days... and now suddenly the rebooting is back....
Last edited:
Sentential said:
Hello, all. Well basically I am a bit of a n00b at this overclocking thing (and system building in general. Anyway I just read this post and you have me worried. I am running a 920 on a P5WD2-e Prem. The first post states you need to plug in all three power conectors. I have the 24pin and 8pin connected, but not the 4pin 'molex' you are talking about.
So my question is this. In think the 4pin molex you are reffering to is labelled in the P5WD2-e Prem manual as 'EZ_PLUG' and it states it only needs to be used when running crossfire. I am not using crossfire so never touched it - however would you recommend i plug it in? If so does it just use a standard molex?
Thnaks all,
Shnorque
