No, the P5Q board isn't inherently bad. Its performance depends on your specific needs and usage.
No, the P5Q board isn't inherently bad. Its performance depends on your specific needs and usage.
I've improved the power supply so voltage levels are stable now. I also upgraded the RAM to extremely dark DDR2, which on this board appears to run around 1250MHz max. Still, no matter what I try, I can't push this board beyond 500MHz FSB—it still handles it well compared to my G31m S2C, which can reach about 540MHz even with lower voltages. Modded BIOS won’t help much either. This setup is already pushing limits, and it’s surprising how a P5Q Pro Turbo BIOS doesn’t improve things.
Back then, not every board was the same. My 790i could handle a very fast FSB, but if you didn’t set it properly, it would fail. You have to adjust each CPU in mind. For my QX9650, the settings won’t fit an E8400, but pushing it around 440 gives great results with a Xeon. RAM speed matters too—like choosing between 1T and 2T—and whether it’s linked or unlinked. If your board supports unlinked, you’ll get much better performance since you’re not limited by RAM constraints. Generally, only the top-tier boards can run unlinked; they’re just different. Each one needs careful tuning and patience during overclocking. I didn’t reach 4.5GHz with my QX using just VCORE and multiplier—it took weeks of adjustments. Even the cross-core voltages on my high-end board were tricky. Cooling is important, but I won’t go into details. What matters is I used a second-hand H100 with a modified bracket to fit 775 and got solid stability, which I wouldn’t have achieved with a tower cooler. Building a custom loop could push it to 5GHz, but you’d need to water-cool the chipset too. The results speak for themselves.
In fact, switching to my p5q instead of my p5q pro gives me a good balance. I can reach around 530 MHz and about 1250 MHz RAM frequency. Probably because the pro model has a damaged ground pin. It’s still a bit disappointing that it only hits 530 when a high-end 64GB board would completely destroy it. The cheapest I’ve found is around $30, while a p5q usually costs about $20. Still, the gigabyte boards are considered legendary for their performance, reaching impressive FSBs. Right now, a P45’s record is around 766 MHz, set by a P45t user. It’s not the absolute top with liquid cooling, but it easily surpasses my p5q in FSB speed. The current P45FSB is about 766 MHz, achieved by someone who probably upgraded it. That’s still way better than what I have. Even my G31M S2C maxes out at 542 MHz, which is impressive considering it has solid cooling and already beats my p5q in FSBs. On top of that, I’ve managed 542 MHz myself.
Otherwise, I’m thinking about swapping the board on that machine. I might sell it and keep a sample for later. I’d need more FSBs to upgrade further, but I could try some voltage adjustments if I can manage it. It would be a hassle, but maybe it’s worth it.
On the other hand, I could go with an X58 or X79 right now. Those are good deals—especially the X58 with a failed BIOS chip. The X79 is $150 for CPU, board, and RAM/GPU. It’s a solid choice if I want something more powerful.
Honestly, there’s another way. Just use the computer I built. That lets me experiment with voltage tweaks, maybe even fix or upgrade it. It could be faster than my dad’s Craptop, especially if I add power delivery and boosters. It’ll be a challenge, but it might work.
I’m also considering selling both models for a gigabyte P45. If I get a good deal, I won’t throw anything away. My parts are likely worth over $100 each—some dead boards, old p5qs, and other gear I might resell. The X79 is $150 for the full kit, while the X58 is $150 plus CPU, RAM, and GPU.
In short, I’m weighing options carefully, balancing cost, performance, and future upgrades.
That’s where the “not created equal” issue shows up. Back then, I had twin Asus 750i boards—one could reach around 590 FSB, while the other barely surpassed 400. That was using simple E8400s. Adding a quad changed everything. Now things get really complicated. I own a mATX P5Q board I just got for a project. I haven’t tried it yet, but I might install an Xeon and run some tests on the FSB limit just for fun. This has reignited my interest, so I’m putting together another 790i with a new loop. I’m hoping to hit 5GHz next weekend. I only managed about 1.5V last time for 4.5, but with enough cooling and pushing around 1.7, I think it could work. You should look for an unlocked Extreme chip on eBay—dual or quad options are affordable now, and shipping from China is manageable if you get it right. The unlocked multiplier makes things much easier. *Edit: Just checked eBay and the QX series prices have skyrocketed. Don’t even consider the dual-core X6800 anymore—it seems nearly impossible to find. What a time we live in! A QX6850 for $80 CAD is the best I’ve seen so far. I think I got my 9650 for under $100 back in 2017.*
I think the boards might not be the same due to physical wear or binning. These low-end ASUs often have BIOS issues, so it doesn’t really matter. I won’t settle for a 65nm chip, even with unlocked multi. Those parts are terrible at handling voltage—even at 4GHz you need around 1.6V, and the danger zone is probably not much higher than 45nm. Maybe closer to 2.1V, which makes them pretty unreliable unless you’re using something like L2. Are you sure you got a 590 MHz on your 750i? The best valid speeds on hwbot are only 530MHz. My cooling is bad, so I can’t really push the CPU faster. I’m planning to build a cheap custom loop soon, and maybe buy some Xeons too—what frequencies can you achieve on your 790i?
I might be confusing the FSB numbers; it was quite some time ago when I had those 750i boards, and they were incredibly high. I haven’t pushed the 790i as far as it could go—maybe I’ll try it soon to find out how it performs. Back in 2010, with an E8400, I managed to get over 500, but it still had plenty of room to grow and I never went beyond that. Cooling was the main concern then, not a big deal now. For those aiming for this path, sticking with 45nm chips seems best. The 65-series feels more like something to explore rather than a practical choice.
Ive also been looking locally for some deals and found 2 broken boards of interest Ep45 ud3p (~5$), seller says that it turns on and off like crazy, prob either a short or they were using a garbo psu P5q pro (~5$) i dont want to use it personally but if i could fix it then i could sell for a profit, seller says that it hasnt been used for awhile and was working before, prob just got dust in the socket and just needs some cleaning Theres also a 3rd board this same seller is selling, p5q se2, seller says that its completely dead so ill have to ask em why its dead Might be able to score myself a valuable good ocing p45 and make a lil profit aswell xD
Just a fun little tidbit. I setup the QX and 790i under a twin 240mm rad water loop to try and hit 5GHz. I'm running into RAM issues at the moment that I'm sorting out, but I have discovered my 790i will do 2000MHz (500MHz) FSB with no issues when pumped up to 1.5v. If I get my RAM issues sorted out with some sticks that the board will actually like then maybe I'll hit 5GHz at around 1.65Vcore. Also I have now successfully pumped 1.8V into a QX9650 and it lived, so... 45nm is pretty durable. Also theoretically if I can pump my FSB to 500MHz (or more even), that means I can get my Xeon X5450 (which has a max multi of 9) to 4.5GHz. I'm going to try that next.
I’m not recommending entering the 1.9V range or using Fsb terms. It sounds like you might be dealing with poor RAM or a faulty system. If you’re building a DIY cooler from scrap metal, that could help boost performance. I noticed your E5800 needed around 1.85V to boot and about 2.1V PLD—why would it require such high voltages?