Page 1 of 7 12345 ... LastLast
Results 1 to 15 of 103
  1. #1
    Joined
    Aug 2004
    Posts
    10,610

    VNF3-250 Temp Mystery Solved

    Please see post 12 if you wish to help.


    There has been a great deal of confusion about the temps reported on the vnf, CO's reading lower than CG's and some CG's reading like CO's. The reason is quite simple, early CG Newcastles are Clawhammer (CO's) with half the cache disabled.

    http://forums.amd.com/index.php?showtopic=17145

    Courtesy Greyhound:
    "Now how do you know what revision your processor has? There are two possibilities:

    If your processor is already installed and you have a running O/S, you can use a CPUID-utility to identify your processor.

    CPUID 'F-4-8' is rev.C0 while CPUID 'F-4-A' is rev.CG - CPUID 'F-C-0' is a Newcastle-part(512KB L2-cache), ALL Newcastles are rev.CG



    "If your processor is not installed, you can determine its revision by looking at the OPN(Ordering Part Number) - its printed on the processor itself.

    Example:
    ADA3200AEP5AP
    CAAKC 0335XPMW

    The 'OPN-suffix' will tell you the type/revision of the core:
    'AP' - > Clawhammer(model 4), rev. C0
    'AR' -> Clawhammer(model 4), rev. CG
    'AX' -> Newcastle(model C), rev. CG(ALL Newcastles are rev CG)

    The figure in red denotes the L2-cache size:
    '3' = 256KB
    '4' = 512KB
    '5' = 1MB

    'ADA3000AEP4AP' - 512KB L2-cache, C0-rev. Clawhammer
    'ADA3000AEP4AX' - 512KB L2-cache, Newcastle(ALL CG)

    'ADA3200AEP5AP' - 1MB L2-cache, C0-rev. Clawhammer
    'ADA3000AEP4AX' - 512KB L2-cache, Newcastle(ALL CG)"

    I think the author typoed the top claw 4ap should be 4ar
    note the "4" and "5" in the clawhammers

    See this amd link for more detail
    http://www.amd.com/us-en/assets/cont...docs/30430.pdf

    The 512KB Clawhammers are MARKETED and CPUIDed as CG Newcastles but their themal diodes are clawhammer, and the diode in the "real" newcastle is different or in a new location.

    Further confusion was caused with the release of the 9-21 bios as it will enable "cool n' quiet" automatically with bios defaults if you have the amd cpu driver installed (from trying to enable c&q with the bios setting in the 7-29 bios) The underclocking and volting will of course lower the idle temps and anything less than actual 100% cpu load. While folding cpu temps are identical.

    Check the forum for people wondering why their cpu's are running @ 1G with the new bios. Answer: c&q automatically enabled.

    Real Temperature. The vnf3-250 with the "real newcastle" has the most accurate measurement of the INTERNAL CPU CORE of any motherboard yet released. If you don't believe, start a new thread and I'll explain my evidence and reasoning.
    Last edited by TheGlasMan; 09-30-2004 at 11:04 PM. Reason: edit: properly place credit

  2. #2
    Joined
    Mar 2003
    Location
    FL
    Posts
    2,753

    Re: VNF3-250 Temp Mystery Solved

    Mine reads 50-57C at 10x240 default vcore so I'm pretty confident that mine is reporting correctly, actually mine seems to report the same temps with all but the 0308/original bios which reported low.
    i7 Beast: Asus P6T Deluxe, Intel i7 920 21x191 (4000) HT on 1.3875v, Swiftech GTZ/D5/BIX2
    6GB (3x2GB) G.Skill pi Black DDR3 1600, ECS 9600GT
    CM STACKER 810, PC Power & Cooling Silencer 750


    i7 Game: Asus P6T, Intel i7 920 20x180 (3600) HT on 1.275v, TRU-120
    6GB (3x2GB) G.Skill DDR3 1333 7-7-7-18, MSI GTX260 OCv2, Dell 3007WFP LCD
    CM STACKER 830, PC Power & Cooling Silencer 750


    AMD: Biostar 790GX A2+, PII X4 940BE 17x200, Freezer 64 Pro
    4GB (4x1GB) Ballistix Tracer DDR2 800, Powercolor 4850
    Corsair VX450

  3. #3
    Joined
    Aug 2004
    Posts
    10,610

    Re: VNF3-250 Temp Mystery Solved

    Nice o/c at stock v, are those temps idle-load, which bios, what is load, what were your temps w/ 3-08 bios? (If you remember.) Your idle temp says newcastle but mine scales a lot higher under load, yours seems to scale like a CO, that tornado must be blasting.

    My later newcastle 3200 is running 2500 at 1.55v, 53c idle, 68 folding.
    I think part of the problem is that the load is not defined and as you can tell a little voltage raises temps a lot. Also my xp-120 dropped temps 10c (load) over my old hs fan tt silent boost lapped.

    I had a paragraph about bios differences, but had a momentary power loss which wiped it out. Darn Tagan psu refuses to be powered by ups. I was to lazy too rewrite it. Yes I agree 3-08 showed much lower and freaked the CO owners out. When I installed the beta bios I almost freaked until I realized it was the bios raising the temps so much.

  4. #4
    Joined
    Jul 2004
    Posts
    115

    Re: VNF3-250 Temp Mystery Solved

    i've got the later newcastle 3000+ and im right now idleing at around 53-55 at 10x230 stock voltage. I can do 10x240 stable easy but cant run this current ram 1:1 at that speed so im running this till i get better ram. I still load at 67-70 or so, which seems to be the internal temp. at 10x240 at 1.7 volts it would load at up to 80/82 or so, but seeing as i've heard that athlon 64s are stable up to 95c on-die i'm not worried. Does this sound about right?

  5. #5
    Joined
    Mar 2003
    Location
    FL
    Posts
    2,753

    Re: VNF3-250 Temp Mystery Solved

    Thanks, I very pleased with the cpu, stepping up to 10x250 though things start to change fast, needs more vcore and really runs hotter than I would like, especially for only another 100mhz.

    It idles around 50-51C and loads at around 56-57C, this is in a non air conditioned FL environment as well. With it being in an open environment its not too bad running the Tornado around 3500ish rpm, full bore is just too much for anything more than tinkering.

    I had a 80mm TT smartfan II on there originally and my temps where in the 60's, so I placed an order for the 92mm Tornado and got a rheobus, like magic the temps dropped. Even with the Tornado turned down to below 3000rpms it still cools better than the TT fan did at full bore.

    My bios is set to default vcore but it overvolts to 1.55-1.57ish, when I had it set to 1.6vcore it used to report right at 1.67-1.7ish vcore in the bios, personally I'm glad it overvolts because undervolting drives me nuts!

    With the 0308 bios it was reporting just about half of what it reports with the later bios's, I can't remember off hand but it was in the 20-30ish C range.
    i7 Beast: Asus P6T Deluxe, Intel i7 920 21x191 (4000) HT on 1.3875v, Swiftech GTZ/D5/BIX2
    6GB (3x2GB) G.Skill pi Black DDR3 1600, ECS 9600GT
    CM STACKER 810, PC Power & Cooling Silencer 750


    i7 Game: Asus P6T, Intel i7 920 20x180 (3600) HT on 1.275v, TRU-120
    6GB (3x2GB) G.Skill DDR3 1333 7-7-7-18, MSI GTX260 OCv2, Dell 3007WFP LCD
    CM STACKER 830, PC Power & Cooling Silencer 750


    AMD: Biostar 790GX A2+, PII X4 940BE 17x200, Freezer 64 Pro
    4GB (4x1GB) Ballistix Tracer DDR2 800, Powercolor 4850
    Corsair VX450

  6. #6
    Joined
    Aug 2004
    Posts
    10,610

    Re: VNF3-250 Temp Mystery Solved

    High Octane , if you can, will you run cpuid (www.cpuid.com) and list the family, model,stepping letters, I suspect you may have a F,4,A Clawhammer CG or my theory is on shaky grounds. Your idle temps may be the result of the high ambiant temps. Regardless that HSF is effective.

    Deadly Donkey, your temps rise under load as I would expect "later" newcastle to. AMD lists 85C in core at stock as when errata will occur. Max Temp(case) is 70C (middle of heat spreader). Rule of thumb is core is 10C higher and edge is 10C cooler. I remind you that all temp readings of the core are approximations. The diodes output is calculated into a temp by an algorithym(sp) in the bios. I ran stable for a month at 78-79C folding but I won't give up my xp120's 10C under load, that was to close to the edge. 80C on my board is 55C at the edge according to my edge thermistor and any error in that measurement is that the edge is hotter.

  7. #7
    Joined
    Jul 2004
    Posts
    115

    Re: VNF3-250 Temp Mystery Solved

    yah i ran stable for around that long with temps up to maybe 82 max folding, but it now folds at less than 70 at default voltage, i can't get it to really boot at all at 10x250 unless i run mem at 2/3 for some stupid reason, since my mem can do the speed it would dof ine. So i dont know what temps i get for 10x250.

    *edit* forgot to mention, it does register as the F-C-0 newcastle, so it is for sure a later newcastle.
    Last edited by DeadlyDonkey; 09-30-2004 at 12:46 AM.

  8. #8
    Joined
    Mar 2003
    Location
    FL
    Posts
    2,753

    Re: VNF3-250 Temp Mystery Solved

    Well I'm pretty sure its a Newcastle, I know it reports as a CG revision no doubt about that. I picked it up from newegg around the end of August begining of September but never bothered to write down the actual stepping from the cpu. From what I understand it would take more juice and effort to get 2.4ghz from a Clawhammer but I don't know for sure.

    I'll find out for sure tomorrow when I'm at the computer, I know for a fact I've never seen temps of 70C or higher.
    i7 Beast: Asus P6T Deluxe, Intel i7 920 21x191 (4000) HT on 1.3875v, Swiftech GTZ/D5/BIX2
    6GB (3x2GB) G.Skill pi Black DDR3 1600, ECS 9600GT
    CM STACKER 810, PC Power & Cooling Silencer 750


    i7 Game: Asus P6T, Intel i7 920 20x180 (3600) HT on 1.275v, TRU-120
    6GB (3x2GB) G.Skill DDR3 1333 7-7-7-18, MSI GTX260 OCv2, Dell 3007WFP LCD
    CM STACKER 830, PC Power & Cooling Silencer 750


    AMD: Biostar 790GX A2+, PII X4 940BE 17x200, Freezer 64 Pro
    4GB (4x1GB) Ballistix Tracer DDR2 800, Powercolor 4850
    Corsair VX450

  9. #9
    Joined
    Aug 2004
    Posts
    313

    Re: VNF3-250 Temp Mystery Solved

    I have one of the 'F-C-0' Newcastles.

    Runs at 9x267 @ 1.55v idle-49 load-61


    -edit- using stock HSF


    -Vi
    Athlon64 (Game Box)
    2800+ @ 2.6GHz (9x290)
    Chaintech VNF3-250
    1GB PQI 3200 TURBO (Samsung TCCD)
    6800NU 16x1 6 @ 375/850
    Custom Water Cooled

    Intel P4 (Workstation)
    2.8C @ 3.4GHz (14x243)
    Abit IC7-MAX3
    1GB Corsair XMS 3700 TWINX (Samsung TCCC)
    9800SE Softmodded Pro @ 410MHz/362MHz
    Custom Water Cooled

  10. #10
    Joined
    Oct 2003
    Age
    38
    Posts
    375

    Re: VNF3-250 Temp Mystery Solved

    I also have a F-C-0, temps in a 67f room read at 37c idle, not sure about load but would estimate around 45c. I have a SLK948 and tornado mounted on top at half speed. CPU is set at 10x220, default vcore (but it tends to range between 1.50-1.58, average around 1.53v). I believe my temps are pretty much right on, in a normal environment (around 75f) temps are usually around 41c idle and 48c load. I am using bios 7/29.
    My Computer: E8400 @ 3.8ghz, Asus P5Q Pro, 4GB pc1066 Mushkin Ram, WD SATA 640GB, BFG OC+ gtx 285, PCP&C 610 Watt, On Board AC97 Sound, Samsung SATA DVD-RW
    Marshall University, Xbox Live: TWolf79, PSN: DH779, XFire: tituswolf, Steam ID: hd779
    Games I'm playing: COD: WAW (PC), TF2 (PC)
    Sharky Extreme 3DMark Team

    Email: hendricks3@marshall.edu

  11. #11
    Joined
    Apr 2004
    Posts
    41

    Re: VNF3-250 Temp Mystery Solved

    Just checked my CPUID again and I have a later newcastle. F-C-0.

    Temps are 65C load and 58C idle. Thats at stock (3400+). Cooled with a 948U and 92 zalman.

  12. #12
    Joined
    Aug 2004
    Posts
    10,610

    Re: VNF3-250 Temp Mystery Solved

    Thanks for all the responses.
    To clarify, all 512Ks are official AMD CG steppings. To avoid confusion of the general public they are all called Newcastle. Only nuts like us care (and then only because of the temp issue ,and overclocking) that there are two versions.

    Maybe we can get to the bottom of this with some defined data.

    CPU:
    Stock speed, stepping (F-4-8, F-4-A, F-C-0) from CPUID ( www.cpuid.com ) family or cpu order code, Clock speed when temps tested, bus and multiplier ( I don't think the mem controller is stressed by these test but you never know)

    Temps:
    Idle and Load from a windows utility. They all show the same temps but sometimes the mix up the outputs, ie chipset temp is reported as cpu temp. I recommend Speedfan www.almico.com/speedfan.php as it shows all sensors, the cpu temp is the one that rises the most when the cpu is loaded (temp 1 in my case). MBM users uncorrected temps please.

    Load:
    If possible can you run F@H, Seti, or Prime95 2nd (In-place large FFTs) torture test for a few minutes till temps stabilize, as I have a good idea what temps I expect. Run with monitoring utility open to be sure you have cpu temp (see above). The Prime95 test will be the most consistant so I would prefer that if possible. http://www.mersenne.org/freesoft.htm

    Voltage:
    Bios setting and windows utility (speed fan has that info).

    I don't mean to sound picky but uniform data is our best chance to get an idea of what our temps actually mean.

    High Octane and ViruX, your responses are putting a big hole in my theory! One with stock cooling the other with aftermarket. I would really appreciate the above data from you guys.

    My board over volts about .03v but remember these readings are ball park, and the error is random, if consistantly on the high side, the actual voltage could be spot on regardless of the reading.
    Last edited by TheGlasMan; 09-30-2004 at 11:08 PM. Reason: added cpuid url

  13. #13
    Joined
    Mar 2003
    Location
    FL
    Posts
    2,753

    Re: VNF3-250 Temp Mystery Solved

    Forgot I had some other apps running in the background that I can't stop yet so I can't go stress testing today, tomorrow I will Prime and get some real up to date full load temps for comparison.

    The cpuid: F-C-0
    Rev: DH7-CG
    Core Speed: 2400.4
    10x240 1:1

    Set to default vcore, reported vcore 1.536-1.568
    Current temp with apps running in background: 53C CPU, 41C System - per Digidoc
    Cpu fan speed: 3668rpm
    i7 Beast: Asus P6T Deluxe, Intel i7 920 21x191 (4000) HT on 1.3875v, Swiftech GTZ/D5/BIX2
    6GB (3x2GB) G.Skill pi Black DDR3 1600, ECS 9600GT
    CM STACKER 810, PC Power & Cooling Silencer 750


    i7 Game: Asus P6T, Intel i7 920 20x180 (3600) HT on 1.275v, TRU-120
    6GB (3x2GB) G.Skill DDR3 1333 7-7-7-18, MSI GTX260 OCv2, Dell 3007WFP LCD
    CM STACKER 830, PC Power & Cooling Silencer 750


    AMD: Biostar 790GX A2+, PII X4 940BE 17x200, Freezer 64 Pro
    4GB (4x1GB) Ballistix Tracer DDR2 800, Powercolor 4850
    Corsair VX450

  14. #14
    Joined
    Aug 2004
    Posts
    10,610

    Re: VNF3-250 Temp Mystery Solved

    Highoctane, I understand completely. When you do get the chance, can you leave the fan speed the same, until stabilized (or as high as you feel comfortable) to see what the temp rise is, (Delta T to idle) before you play around with the rheostat. It's fun even with the low speed fan I'm using.

    The xp-120 allowed me to run at 2500 at 1.55v (1.6v indicated) down from 1.6 (1.67 indicated) which helped the temps even more. it seems the gap expands at higher V. Like you I'm happy, but maybe someday I'll get a powerful 120mm fan and see how fast this cpu can go.

  15. #15
    Joined
    Dec 2002
    Location
    Blaine, WA
    Age
    64
    Posts
    30,989

    Re: VNF3-250 Temp Mystery Solved

    Thanks TheGlasMan, a very informative series of posts. What you have done has gone a long way to help explain the differences we see in reported temps. It seems that the motherboards aren't specifically at fault, nor are the BIOS's per se.

    I have been actively looking for good data to pass on the Chaintech engineers and this post will certainly be included, as it goes a long way to sorting out many of the erroneous readings.

    Me? You got me....I'm going to try Speed Fan (for the first time)...been with MBM forever (it seems) but it is on it's way out as it isn't actively supported anymore.
    Main Rig: EVGA 141-BL-E769-A1 LGA 1366 Intel X58 CLASSIFIED/WATERCOOLED ED. cooled by the Monster Water Cooling Setup
    HTPC One BIG Case Asus 8 CORE Monster
    ASUS: 5 * BIOSTAR: 1 * CHAINTECH: 1 * EVGA: 3 * GIGABYTE: 5 * SUPER MICRO: 1 *TYAN: 2
    180+ GHZ total power for PC Perspective Killer Frogs Rosetta @ Home Team as The Uncle B's!!




    Spoiler!

    I'm the Uncle your Aunt won't talk about. Go ahead and pull my finger!

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •