|
|
Post by MachineLearning on Jun 7, 2022 14:41:17 GMT -5
Hi all, I've always found nVidia chipsets to be intriguing, mainly due to them being avoided by most enthusiasts post-NF-4. Overclocking is generally worse than their Intel / AMD counterparts, although the 780a & 790i aren't too bad in my experience. I have quite a few nForce chipsets, and either nobody puts their best effort into OC-ing them (likely), or I have a knack for nForce. Either way, here are a bunch of nVidia mobo ref freq scores  370.2 MHz - 980a (ASUS M4N98TD Evo) hwbot.org/submission/5264123_machinelearning_reference_frequency_m4n98td_evo_370.2_mhz#2 for 980a, #1 for board. Seems to be board limit. 429.8 MHz - 780a (Foxconn Destroyer) hwbot.org/submission/4989433_machinelearning_reference_frequency_destroyer_429.8_mhz#4 for 780a, #1 for board
529.95 MHz - 750i (EVGA 750i SLI FTW)
#1 for board, #2 for 750i
558.98 MHz - 750i (XFX 750i SLI) #1 for 750i. Crossflashed to MSI P7N Platinum BIOS, it's leagues better 545.63 MHz - 780i (EVGA 780i FTW) #1 for board, #5 for 780i
573.91 MHz - 790i SLI (ASUS Striker II NSE) #1 for board, #6 for 790i SLI
369.98 MHz - GeForce 8200 (Jetway NC62K-LF) hwbot.org/submission/4995731_machinelearning_reference_frequency_nc62k_(nf_mcp78)_369.98_mhz#1 for GeForce 8200 460.04 MHz - GeForce 9400 (DFI LP JR GF9400-T2RS) hwbot.org/submission/5634203#1 for GeForce 9400. This board is a buggy mess and I couldn't recommend it for anyone even if its performance was good. 435.92 MHz - 590 SLI (ASUS ROG Crosshair I) hwbot.org/submission/5132168_#1 for board, #2 for 590 SLI
450.31 MHz - NF4 SLI (DFI LP UT NF4 SLI-DR) #4 for board, #17 NF4 SLI
Work in progress. Multiple CPUs which wall at about 450-451MHz. Goal is 500MHz
389.06 MHz - NF4 SLI x16 (ASUS A8N32-SLI Deluxe) #2 for motherboard, working on more CPUs
370.12 MHz - NF4-4x (MSI K8N Neo3-F) #2 for board, this is my only s754 CPU for now so I need more samples to know if it's board limit.
254.36 MHz - nForce Professional 2200 (ASUS K8N-DRE) #1 for board; need to voltmod!
[Work in Progress] 239.80 MHz - NF2 Ultra-400 (Abit AN7)
#13 for board, need more samples
|
|
|
|
Post by MachineLearning on Aug 18, 2022 10:36:02 GMT -5
|
|
|
|
Post by MachineLearning on Aug 26, 2022 16:59:43 GMT -5
New top XFX 750i (MB-750I-72P9) FSB score, 517.42MHz: hwbot.org/submission/5068133_Board's vCore maxes out at +0.3v offset... so max for my CPU ended up being ~1.37v under load - impressively low ceiling. |
|
|
|
Post by Macsbeach98 on Aug 26, 2022 18:12:23 GMT -5
These boards do run Quads no problem?
|
|
|
|
Post by MachineLearning on Aug 26, 2022 18:23:22 GMT -5
|
|
|
|
Post by freeagent on Aug 26, 2022 21:14:23 GMT -5
I had a 680i, and for whatever reason, it just hated my Q6600. Could barely run the system @ stock, was not stable. It did ok with my 6300 though. I haven't touched an Nvidia chipset since.. But I did have awesome luck with NFII and NF4, and later Intel  |
|
|
|
Post by MachineLearning on Aug 26, 2022 21:26:30 GMT -5
I had a 680i, and for whatever reason, it just hated my Q6600. Could barely run the system @ stock, was not stable. It did ok with my 6300 though. I haven't touched an Nvidia chipset since.. But I did have awesome luck with NFII and NF4, and later Intel C2Q on the 600-series was totally busted until mid-lifecycle. You were not alone www.anandtech.com/show/2198/5Apparently it took a minor hardware revision for some boards, BIOS update for others (Striker). I'll be getting around to NF4 & NF2 eventually too  |
|
|
|
Post by Bones on Aug 26, 2022 21:41:23 GMT -5
I had a 680i, and for whatever reason, it just hated my Q6600. Could barely run the system @ stock, was not stable. It did ok with my 6300 though. I haven't touched an Nvidia chipset since.. But I did have awesome luck with NFII and NF4, and later Intel C2Q on the 600-series was totally busted until mid-lifecycle. You were not alone www.anandtech.com/show/2198/5Apparently it took a minor hardware revision for some boards, BIOS update for others (Striker). I'll be getting around to NF4 & NF2 eventually too You can try to beat one of mine while you're at it - Absolute WR for Socket 754.  hwbot.org/submission/4245674_bones_reference_frequency_infinity_nf4x_495.77_mhz |
|
|
|
Post by MachineLearning on Aug 26, 2022 21:44:11 GMT -5
Now that is dope |
|
|
|
Post by MachineLearning on Aug 29, 2022 14:25:06 GMT -5
Just tried all my E0 E8500 CPUs on the SIIE. There's a hard wall at 550MHz FSB, none of them are able to boot any higher. Even with (far too high) VTT up to 1.5v, SPP up to 1.7v. E6750 couldn't do 525, E7500 can't do 500. Tried adjusting GTLs in every direction, disabling second core, slotted GPU in the MCP PCIe, SPP PCIe @ 160MHz, LDT @ 3x200, all with no luck. I'm totally confused as to how Christian Ney & Pavo managed to get any higher: hwbot.org/benchmark/reference_frequency/rankings?hardwareTypeId=motherboard_4562&cores=#start=0#interval=20 CN using the same BIOS as me. Pavo I can't tell since it's legacy validation.
And in the process of using my Thermalright TBK (bolt-through for many sockets) for the SIIE I hurt a capacitor.  I didn't realize it was directly under a too-long screw. Don't know if I should be worried long-term.
Totally at a loss for > 550MHz FSB  |
|
|
|
Post by Macsbeach98 on Aug 29, 2022 17:37:17 GMT -5
I would replace the capacitor even though its probably working ok at the moment.
|
|
|
|
Post by MachineLearning on Aug 29, 2022 17:56:49 GMT -5
I would replace the capacitor even though its probably working ok at the moment. Welp, if I'm gonna replace one and order some caps I might as well do it better. And maybe the whole VRM output while I'm at it... Mismatched caps in the same line will bug me. I gotta practice my soldering first. Think there are cap pads on the bottom of the socket too. Gives me ideas |
|
|
|
Post by MachineLearning on Nov 28, 2022 22:17:48 GMT -5
|
|
|
|
Post by n2gaming on Dec 3, 2022 22:01:45 GMT -5
I am a NForce chipset enthusiast. Own several from 939 to AM3. I might have one or two earlier nforce based boards but I’d have to dig them out and get a model number to see what I have. Below is a list of working Nforce boards I have left after selling a couple about a decade ago.
1x DFI NF4 SLI DR
1x ASUS M2N32 SLI DELUXE
3x FOXCONN DESTROYER
2x MSI NF980 G65
CPU prices are fairly decent lately so I have started stocking up for making some on air runs.
|
|
|
|
Post by MachineLearning on Dec 3, 2022 23:27:36 GMT -5
I am a NForce chipset enthusiast. Own several from 939 to AM3. I might have one or two earlier nforce based boards but I’d have to dig them out and get a model number to see what I have. Below is a list of working Nforce boards I have left after selling a couple about a decade ago. 1x DFI NF4 SLI DR 1x ASUS M2N32 SLI DELUXE 3x FOXCONN DESTROYER 2x MSI NF980 G65 CPU prices are fairly decent lately so I have started stocking up for making some on air runs. I'd call myself the same. Just got a few more NForces this past month: NF2 Ultra 400, NF4-4x (s754) and NF4 SLI (s939). When I get familiar with each platform and get proper CPUs for raw BCLK, they will end up here are well. Nice MSI 980a, that's the only DDR3 980a with video out as far as I'm aware. I adore the Foxconn Destroyer, although it's shit for Phenom OG dual-cores. I mentioned this on another post here recently, all their high-end boards are beautiful to me... My preference is niche brands although I have way too much ASUS to really say much more. If you need BIOSes for the Destroyers, I have many of them as should Scotty. EDIT: Wait, I thought your name looked familiar... You were all over those Destroyer owner's threads. I remember reading all of that just to scrape the files I've got. So I suppose you need nothing from me... |
|
|
|
Post by n2gaming on Dec 4, 2022 0:34:41 GMT -5
I don’t know if I still have or where any of the bios files I had from way back then would even be today.
|
|
|
|
Post by MachineLearning on Dec 4, 2022 5:42:25 GMT -5
|
|
|
|
Post by MachineLearning on Dec 31, 2022 0:55:35 GMT -5
|
|
|
|
Post by somerandomtechyboi on Jan 5, 2023 9:47:38 GMT -5
New top XFX 750i (MB-750I-72P9) FSB score, 517.42MHz: hwbot.org/submission/5068133_Board's vCore maxes out at +0.3v offset... so max for my CPU ended up being ~1.37v under load - impressively low ceiling. Havent tried pad modding? think its basically just set auto volt in bios (iirc 1.2v is default for most 45nm) and bridge pads based on the voltage you wanna set, 1.6v is 0 (grounded) on all 8 vid pads so if you set auto volt then its a matter of just bridging vid 6 and 2 to vss, if you need to set a pad from 0 -> 1 then bridge to vcc. Vid tables should be in pretty much any 775 cpu datasheet, to find the page its just look at the contents index thingy at the start of the document, go to the tables section, and find a table thats something along the lines of voltage identification definition |
|
|
|
Post by Macsbeach98 on Jan 5, 2023 17:32:03 GMT -5
Just put a trimmer on works with all CPUs then.
Pad modding is better for BSEL mods.
|
|
|
|
Post by somerandomtechyboi on Jan 5, 2023 23:33:04 GMT -5
Just put a trimmer on works with all CPUs then. Pad modding is better for BSEL mods. Ofc real vmod is better, not limited to a meager 1.6v so better for xoc and not just dailying or mild oc, but pad mod is decent if you cant solder a trimmer onto the vrm actually im kinda curious as to how far you can go with proper vmod cause i know 2v should be attainable most of the time but what about going higher than that? Sure most cpus will problably instadie >2.2v anyways but how far can it go? |
|
|
|
Post by Macsbeach98 on Jan 6, 2023 1:57:57 GMT -5
That depends on the controller some have OVP protection built in of .15v over so if the vids go to 1.6v you will only get 1.75v and it will shut off.
Others will go over 2v easily.
|
|
|
|
Post by somerandomtechyboi on Jan 6, 2023 6:10:47 GMT -5
That depends on the controller some have OVP protection built in of .15v over so if the vids go to 1.6v you will only get 1.75v and it will shut off. Others will go over 2v easily. Any way to disable ovp? My p5q is retarded and dont wanna go past 1.75v on 45nm dualcores or itll just roll to d5 code and shutoff, not even jumping ps_on to gnd will force it to continue post. But i think i have pushed a p4 631 to ~1.8v on a random bios run to 5.6g (now dead cause chipped the cpu while scraping off the indium after delid, it was degraded anyways) And speaking of cheddar mill cpus what boards can run a celeron d 65nm? My p5q just 00 and doesnt wanna post same with giga ep45, dont think ive tested with my giga g31 but eh i think mines been dead for awhile since i chipped it cause it fell onto a corner. Im looking at some cheap giga p31 or p35 boards to run celeron d >5g cause i gotta do something for fun every once in awhile and not just constantly fixing stuff |
|
|
|
Post by Macsbeach98 on Jan 6, 2023 7:11:46 GMT -5
P5Q you should be able to trimmer mod it 1.7v would be the high voltage for it now.
You wont run Celeron or Pentium in Gigabyte P45 you can get them to boot but you cant adjust voltage they are VRD1.0 Cpu's have to be VRD1.1 Cpu's.
You can run them in Giga P35 though the magic board for celerons is Asus P5E3 premium X48 good luck on finding one of those though.
|
|
|
|
Post by somerandomtechyboi on Jan 6, 2023 8:49:52 GMT -5
P5Q you should be able to trimmer mod it 1.7v would be the high voltage for it now. You wont run Celeron or Pentium in Gigabyte P45 you can get them to boot but you cant adjust voltage they are VRD1.0 Cpu's have to be VRD1.1 Cpu's. You can run them in Giga P35 though the magic board for celerons is Asus P5E3 premium X48 good luck on finding one of those though. Not even with trimmer mod you cant adjust voltage? Yea no x38 or x48 boards here in indo, very rare and most of em are problably garbage from gigabyte and not p5e though giga p35 dont seem to have vpll adjustment and ocing past 5g on the p5q with a p4 631 needed vpll to be raised, on the 5.6g run i just shoved 2.4v vpll and yea makes sense why it degraded. Kinda curious on the giga p31 though cause i can get a cheap p31 ds3l at 8$, p35 s3l ~7$, though it seems like theyll clock the same anyways, p31 and p35 ds3l same board layout holy grail for g chipset overclocking would problably be giga g33 cause apparently someone at hwbot managed to get one near 600fsb but alas no giga g33s here either there is some guy selling a giga g45 ud2h alongside a cpu and some ram at an absurd price, maybe i can take only the board and buy for a more reasonable price, its not an asus so no garbage bios crippling and g45 isnt walled to 440fsb like the 43 chipsets atleast looking at hwbot. Though i gotta wonder if the giga g33m ds2r just has a really good board layout if it can run near 600fsb on a non p35 chipset |
|
|
|
Post by MachineLearning on Feb 9, 2023 2:48:44 GMT -5
These boards do run Quads no problem? I was playing with a random Q9400 today with the SIINSE crossflashed to SIIE and was thinking of this question. I have not pushed an E0 quad but I intend on trying it out and seeing if it fares better than R0. Board does at least 500MHz w/ R0, beats my DX38BT w/ Q9550 by 20MHz. For other examples on the Bot I manually searched for any reference frequency with nForce chipset and "core 2 q" in the CPU name. There are 2 results for 790i > 520MHz, both on ambient. Whereas X38 w/ same criteria, the highest listed is 507MHz. X48, there are 3 > 520MHz, two on ambient and one on dice (545MHz). Granted this won't account for people who ran benches at a higher FSB but never submitted a reference frequency. 790i is doing way worse than P45 on quads for sure but it's still an okay all-rounder chipset IMO. If you ever got hold of one I would be very eager to see what you can get out of it. |
|
|
|
Post by Macsbeach98 on Feb 9, 2023 16:59:19 GMT -5
P45 is the best for Quads Gigabyte especially, X48 for duals It would have to be 600 bus capable and have good efficiency before I would persue one somehow I dont think you will ever see me with one.  |
|
|
|
Post by MachineLearning on Mar 16, 2023 6:54:48 GMT -5
XFX 750i is a near clone of the MSI P7N Platinum, and takes a crossflash just fine with the newest AFUDOS.EXE. From that I was able to get a new high for 750i: hwbot.org/submission/5227220_Before this I was under the assumption nVidia gimped FSB OC for this chipset to around 525MHz but it seems that isn't the case. 539.8MHz is decent 780i territory. The MSI BIOS adds OC profiles, better FSB (+20MHz) and memory OC, and eliminates the need for the XOC Jumper. Also, BIOS v1.0 has higher adjustment ranges for vDIMM (3.1v vs 2.5v) and vMCH (1.59v vs 1.51v). |
|
|
|
Post by somerandomtechyboi on Mar 17, 2023 20:54:58 GMT -5
didnt try flashing a 790i bios on it?
Better fsb from crossflash seems to be a result of better ebb so look for a 790i ebb but best one ive found is the maximus ii formula but its p45 and i have no idea if a p45 ebb will work on an nvidia chipset, i have gotten a commando, p5k-e (or was it premium) and REX ebb working on my p5q but unless you got a hot air tool or something to externally flash bios then wouldnt reccomend trying a maximus ii formula ebb
For ebb swap its as simple as finding mmtool bkmod fix on the internet and downloading the 2 bioses
|
|
|
|
Post by MachineLearning on Mar 26, 2023 17:53:22 GMT -5
I've had this MSI K8N Neo3-F board for a few months but just got around to it. It's really easy to hit high FSBs with but the vcore will need to be modded for serious OC. Only VID up to 1.55v set but at least it ends up 1.59-1.62 generally. Managed 370.12MHz with my Clawhammer, I'm sure it can go higher after I buy some more 754 samples. hwbot.org/submission/5236432_It's not such a bad board, but without a vCore mod I can't quite recommend it for people crunching out s754. It's okay for now though. |
|
just barely!
I didn't realize it was directly under a too-long screw. Don't know if I should be worried long-term.
