Jump to content


  • Content Count

  • Joined

  • Last visited

About nYdGeo

  • Rank
    New Member
  1. I have some experience that supports most everything covered here. I have an old Clawhammer FX55 that on a given system was reaching 61C under P95-induced load. I removed the IHS, and noticed that the Thermal paste appeared to have an air bubble in it, and generally looked like a crappy job. I cleaned things up, applied a grain-of-rice worth of AS5 on the top of the core, replaced the IHS, and reassembled the system. After a couple days, the P95-induced load temps were now 52C. I removed the IHS, cleaned things again, applied the same amount of AS5 to the core, and reassembled the system yet again. After a couple of days, the load temps were now 50C. Generally speaking, I'd say if your CPU runs at what seems like appropriate temps for the conditions it operates under, don't uncap it. If however you have one of those CPUs that obviously runs much hotter at a given voltage, speed, and using like cooling to others that post their results, then it may be a realistic option. Since I am using a $100, socket -939 Opty 170, and replacing it is no big deal, I'd say that the warranty issue is a non-issue. However, if you have the latest bad-boy FX CPU, you might want to think twice. I hope that helps someone. Thank you for your time.
  2. If you read this thread from the beginning, you'll see brilliant observation, brilliant thought, top-notch reasoning ability, and maybe even become more impressed with your fellow man. The down-side is that you'll also take in enough ignorance that it may adversely affect your intelligence. Some of it is simply unrealistic; some of it observantly short-sighted, while still more of it is just sheepishly compliant. Now and then it just reads as though it was bought 'n' paid for. Ever notice how most of the folks that make offensive, inflammatory statements calling other "whiners", etc, or say completely unrealistic things like, "If you don't like it don't buy it," are so often the folks that cannot even form a proper sentence? Perhaps that ought to be a forum rule; if commas and periods are too complicated for you, read more and write less. All pokes and digs aside, yes, there has been much noise with the introduction of each new MS Operating System, but nothing like what is being seen now. After reviewing the Vista/Longhorn platform, the United States FAA will be instituting its 45,000 new PC's this year with SUSE Linux...servers and workstations. Another federal office is also moving to SUSE with an estimated 15,000 workstations this year. Many if not all national government offices in the country of France are moving from MS to Ubuntu. For the record, I am not anti-Windows; I like Windows XP though it is not perfect. I also like Linux, though it is not perfect either. I am, however, pretty much anti-Microsoft, in spite of the fact that I give them the credit that they are due for their huge part in shaping modern computing, and even our world itself. If the companies that most any of us work for had as many felony convictions as MS has they'd be out of business. MS is one of the handfuls of corporations in existence that can afford to survive the huge fines and legal fees they have paid. But let me get back on track. I know that even some Linux users keep stating that Linux is not currently a substitute for XP, but I disagree. Yes there can be some pretty serious limits brought on by application availability, or lack of availability. Also, although I believe that any computer-savvy person that is willing to apply a reasonable amount of effort can pick up installing and configuring Linux readily, I will concede that it is not as easy a process as it is in Windows. I believe that this is largely due to familiarity; the average PC user of today would struggle just as much if faced with learning to use DOS & 3.11. The Modern Linux desktop is light-years beyond that. From a performance standpoint, with the exception of (mostly DirectX) gaming, my experience is that modern Linux desktop OS's are also faster, more stable, and much more powerful that XP could possibly be. I also play all of my OpenGL-based games in x86_64 Linux, the newer ones with SLI support; they play without issue, and with plenty of eye-candy. I thank NVidia here for their driver support. This is getting much too long, so I'll cut to the chase. With the exception of an extra learning curve, and application availability, I don't think one has to give up much to live with Linux. Some additional driver support from the hardware industry would help as well. But for every one of those developers that work in Unix, etc, there are a thousand working in Windows, by necessity. And yes, wanting and needed to have a job makes it a necessity. Saying 'if you don't like it don't use it' or 'don't buy it' is an unrealistic argument. Whether it is good or bad, right or wrong, and whether I or anyone else likes it or not, most PC-related jobs involve working with Windows and/or Windows networks. Because of all of this, there is currently and probably always be more applications for Windows. I have many Windows apps working in Linux, but other will not. So again, by necessity, whether it is good or bad, right or wrong, and whether I or anyone else likes it or not, I must keep a Windows installation in order to do certain things in this life. It is not just a choice to use of not to use an OS; it becomes choosing whether or not you want/need to or do not want/need to do things. To those that continuously attempt to solve the MS/Vista debate in a few flames, let's not pretend that this is simple. It is not. Most of the folks here are reasonable and rational enough to know that many of us will eventually have a Vista installation whether we like it or not...by necessity. Hell, I actually like the Aero interface, and love the fact that DirectX 10 will finally, really use our graphics hardware...like OpenGL always has. I disparately want to like the next version of Windows. But none of this changes what MS has done with Vista, or the people-control that they are attempting to institute. On certain levels, some of it is almost frightening, but its just wrong all over the place. Thank you for your time. Let the bon-fires begin! (This post checked for spelling and grammar using “OpenOffice Writer v2.2”)
  3. OKay, cool, thank you! That it awesome, and makes sense since the stock BIOS default settings often won't boot with two sticks of RAM in the board. As for Merlin's BIOS, I'm using it now and it works great. It is indeed a modded BIOS, as even the BIOS screens are different. I am trying currently to learn more about what changes he made. Also, he only made them for the Expert model. Currently, our system with the non-expert SLI-DR board has the factory 406 BIOS in it. Oddly, since I upgraded to this BIOS on that system, the trusty 1GB OCZ Plat Rev 2 memory that I used forever at 250MHz using 2.5-3-3-8 now runs perfect only at Default speeds. It will run 2-2-2-8 @ 200MHz at any one of a number of settings that I've found. Currently its running using slightly modified 200MHz settings by AndyT, and will pass memtest, as well as in-Windows memory testing (Orthos memory testing, etc), or anything else that we throw at it including heavy gaming. Edited 9-23-05 by AndyT DFI NF4 (all), Venice, San Diego or Winchester 90nm and OCZ TCCD/TCC5 memory, 200mhz memory clock 1:1 with HTT speed. In DRAM Configuration Section of Genie BIOS: 200 enable 2.0 02 07 or 8 02 07 or 8 10 16 or 14 03 or 02 03 02 or 01 03 or 02 3120, 4708, 0648, or 2560 (4708 works well at high overclocks) auto enabled auto 0 level 6, 5 or 7 auto DRAM Response Time: fast or normal (as needed if your BIOS has this feature.) auto 256 disable 16 07 disable In Genie Bios: 200 auto 16 16 auto 100 startup 1.35v above VID * Auto 1.30v 1.70v RAM-2.8v (2.7v-3.0v may work for you, depending on final speed) However, using AndyT's 250MHz settings or even his 275MHz settings, while it will pass memtest, it will fail any in-Windows memory test at even 218MHz (lowest I tried over 200MHz). The system is unstable, spontaneously reboots, etc. I tried re-flashing followed by a long clearing of the BIOS, but got the same results. I am truly puzzled by this, which explains why I am looking into the whole BIOS thing. I have read many things here and on other forums stating that when something like this makes little to no sense, it is often the BIOS causing the problem. That got my curiosity up. If anyone has any constructive thoughts here, I'm all ears. Also, still trying to find someone that has compared Merlin's latest against the factory 406, the later BTA's, etc. I wish that I had the time to do it myself! If anyone has done anything like this, as I said I'm all ears. Thanks again folks!
  4. I recently read that Merlin's Expert BIOS's are geared towards TCCD and UCCC based memory. I don't know if this is correct of not, but I can say that his latest works great with my CPU to well over 300MHz, and my OCZ 4000 EB running ratios. Anyway, can anyone tell me what the changes to Tony's 406 are for? The reason that I ask is that Merlin's BIOS is not available for the non-expert versions of the board(s), so I'm looking for the best possible BIOS for the non-expert SLI-DR. The factory 406 works great, and just want to know if Tony's 406 version would be worth trying on that system, with that type of RAM, etc. Thank you!!
  5. True, the temps seem okay either way, but I have somewhat of an OCD about accuracy. That is very interesting info, Praz, and may explain my observations. Do you by chance have any details as to what is going on while we're in the BIOS? I'm just trying to understand why it would be at 60% load. However, that fact would indeed explain the differences, even equalizing the BIOS temps and the temps in Windows! So everything could be off equally by 'x' number of degrees, everything could be spot on (maybe unlikely), and except for my OCD, who cares? If anyone else knows any specifics, they would be nice to know. For now, knowing that the CPU is NOT idling while in the BIOS explains much. Thanks for your time and the info!
  6. I was recently told that the DFI LP UT NF4 Expert model motherboard CPU Temp sensor reports temps in Windows that are about 10 degrees lower than the accurate temps. I went into the BIOS, which I assume is as close to absolute idle as possible, and it shows 25C in my current configuration. Idling in Windows, MBM5, NVidia Monitor, etc, all report 19C, or 6 degrees lower. Assuming that the processing running in Windows raise the temp 3-4 degrees above the temps reflected in the BIOS, that 10 degree thing seems plausible. I can always tell MBM5 to compensate by 'x' degrees to make it more accurate, but I lack accurate info as well. Can anyone shed some factual light on this issue? Also, if this is correct (or close to it), is this also correct for the non-Expert versions of the same board? Thank you in advance for your time and assistance!
  7. Look into the single 120mm WC Kit from Swiftech. It has plenty of capacity, and should be able to be replace the 120mm in the back of the case. If not, as was pointed out already, their Radbox will allow you to hang it off of the back of the case; not the best looking, but its all about funtion herre, and this functions well. I had a Thermaltake Big Water for a couple of years when they were brand new; since it ran hotter than premium water-cooing units, it seemed to lose more liquid to vapor, etc, requiring more maintenance. For the record, I am not complaining; that unit served me well. However, having a more powerful unit, and that Swiftech is much more powerful, not only cools the CPU to lower temps, but also in my experience needs less attention, less refilling, etc. Understand that to some of us, having to do anything to our PC within a 6-month time period is considered a hassle. To others that might live in dusty area such as myself, or who just enjoy babying our PCs, and who clean air filters each week, blow out the case monthly, etc, the adding of blowing out the radiator (to blowing out the case), or adding an ounce of coolant every few months is a completely acceptable amount of labor to enjoy the quiet, effective cooling solution. Since I am currently only cooling my CPU with my liquid cooling, there are times that I wish that I did have one of the extremely nice Freezome units! They aren't quite as effective as a hard-core TEC system, but one heck of a lot easier to live with! But, since I have such an overkill liquid setup, all I need to do is waterblocks and I can also easily cool my chipset, and both video cards with barely any rise in CPU temps at all. Still, a Freezone can cool (up to) a (dual-core) CPU much more effectively that a liquid solution can, and it should be maintenance free; the better unit is rather expensive, but certainly not any more expensive than some premium liquid solutions, and it'll fit in most any case, too. If you have the cash, don't mind the very complicated installation, and want to get way lower temps without the ordeal of phase-change, you could get a standard TEC system. Like the Freezone it only cools the CPU, and its even more expensive, but offers ridiculous performance! On the other hand, a top-end liquid solution cannot bring CPU temps as low as a Freezone can, but ultimately has a much higher cooling capacity. This makes the premium liquid solution a great choice if you'd like to cool more devices. It is also a better choice than the Freezone if you want to cool a highly overclocked (3.5GHz+) quad-core CPU due to the tremendous heat-wattage that these produce; note that the standard TEC unit should also handle that very well. There is some compromise to any solution, so just find the one that best meets your needs. Oh, and secretly I do fantasize about adding a TEC to my system! Something about 5-7C under full load just sounds...cool?
  8. Good day! Since I'm stuck in dual-core AMD-ville until the end of the year, I just replaced my 4400X2 with the Opty 170 in my sig. I'm only using default voltage for the first couple weeks, heat-cycling for the first week for the AS5. I did a quick'n'dirty with Systool and got an error at 2831MHz, so I backed off to 2.75GHz and ran Orthos for about 3 hours. It was time for a cooling cycle, so I shut down then. For the Opty folks out there, is 2.75GHz decent for an Opty 170 at default voltage? If anyone is familiar with this stepping and has tested with it, can you tell me if I have a shot a 3GHz later, or is this a stepping that doesn't respond well to voltage, etc? Any info is much appreciated! Also, can anyone suggest a safe yet productive maximum core voltage for me to use based on the temps that I'm getting? Thanks again! You folks are a wealth of info!
  9. What a disappointment the official support in this forum is. The last time I was here I ended up finally asking whether I should RMA the board. I asked specifically for Angry's input. I never got it, although many of the regulars contributed, and I thank you all. I finally made the decision myself and got a new board. This one seems to work. Here is the skinny, something that obviously no one knows, including the DFI employees...or like last time, they simply never bothered to reply. If you install one of these boards with SLI enabled, youcan go to a single card and back again. But if you install it withone card, YOU MUST REINSTALL THE CHIPSET DRIVERS TO GO TO A SLI SETUP. Then you can install the video drivers and get on with your life. I am sure tired of being the heavy here. The documentation is terrible. I got great response from the forum regulars, and nothing from the actual DFI employees. So, there it is.
  10. I read the huge write up on dividers, RAM speed, etc, and I think that saying it is pointless is incorrect. It does indeed have a MUCH smaller affect than it would say on a Pentium machine of course. However, comparing numbers using the same speeds achieved with and without dividers using benchmarks like 3DMark'05 is misleading, as the CPU has so little to do with it scores in 3Dmark'05. It was by design (and maybe foolishly) created to isolate the video card. Since games have now gotten so (relatively) CPU intensive again, '06 has added the CPU back into the equation. With my old X800 Pro (at 570/570) I tested on my old DFI NF2 UI-based machine at 2.5Ghz (250x10) and with my fiance's old HP (ASUS NF2) machine running a 2200+ at stock speeds; the 3DMark'05 scores were 5965 & 5811. The difference in 3Dmark'01, which is far more CPU based, was more dramatic, with scores of 23,455 vs 13,981. I know that is an exagerated example, but if you go back over the article about RAM speeds on A-64's, notice that there is virtually no difference in '05 scores, but also that there is a small, yet beyond margin of error, noticable difference in '01 scores. I have run both of the A-64 machines here both ways, and I have seen distinct increases in performance in both 3Dmark'01 and '06 using higher memory/bus speeds. It also has the side benefit of running the CPU slightly cooler at the same speeds. I believe that the main point of the article was to make clear that you do not need to use high RAM speeds with an A-64 system to enjoy extreme performance, that it makes much less difference than it does on other types of systems (Pentium, Athlon XP+, etc), and because of this you can enjoy really high performance using even modest grade RAM. Thank you for your time, and I hope that this helps clear things up for someone.
  11. This was simple; fiance's machine: DFI LanParty NF4 Ultra-SLI-DR (Tony's 704-2bta Bios, NF 6.70 drivers) AMD-64 4000+ San [email protected] (250x11, using the HSF from my FX) 2x512MB OCZ Platinum Rev.2 PC-3200 @250Mhz, 2.5-3-3-8 (2.8vt) 1x 160GB WD SATA-II HDD LiteOn 16x DVD-R/RW 2 x eVGA 7800GT rev.517 in SLI (Forceware 84.21) Onboard Realtek AC’97 Audio (Driver A3751[]) Aspire 520wt. PSU Windows XP Pro SP 2 In an attempt to resolve another issue, I: 1) Flashed from a completely stable 704-2bta to 11/14 BIOS 2) Reloaded BIOS 3) Failed first iteration in SuperPI 4) Crashed in 3Dmark'06 5) Flashed back to 704-2bta BIOS 6) 32MB SuperPI in 26.04 7) Passes every stability test Moral: If it works...you should probably thank Tony!
  12. Downloaded and tried the official 11.14.05 BIOS, but still don't see the option to enable anything resembling SLI. So far the results are the exact same. Can anyone tell me if this otion is in this BIOS revision, and if so, where it is? This absolutely blows chunks. This is the second board of this type (non-expert) and have not been able to get either one to do what it's designed to do. I believe in the product, but damn, is it supposed to be a pit of frustration?
  13. in your sig, under NF4 SLI-DR it has the date 11/14/2005. I assumed that was the BIOS revision that you are using in your board. I just learned that there was a BIOS from 11/2005 and was wondering if that is what you are using, is it working well for you?
  14. How is that 11.14 BIOS? I didn't even know that it existed till a few minutes ago. If it has the option maybe I'll try it.
  15. I think that I remember seeing something like that, except maybe 6600, on the Genie screen maybe, using the factory BIOS. But I do not see the option at all using Tony's 704-2bta BIOS. Was this BIOS not designed for SLI or did I miss that optionon another screen? Maybe I should try the latest factory BIOS...hmmm.
  • Create New...