Overclock i7-920 or upgrade CPU?

[loulis]

Hello!

I would like to raise the performance of my old machine and I would like your advice.

The specs of my PC:

Motherboard: Gigabyte EX58-UD5 (Socket 1366) rev1.0 (BIOS F13)
CPU: Intel i7 920 @ 2.67GHz
RAM: 18GB DDR3 @ 1600MHz (9-9-9-24) [ 3 modules of 2GB Corsair CM3X2G1600C9 & 3 modules of 4GB Corsair CMX4GX3M1A1600C9 ]
Operating System: Microsoft Windows 7 Ultimate (x64) SP1 Build 7601
GPU: Gainward GeForce 9400 GT
PSU: 625W Enermax MODU82+ EMD625AWT
Disk drives:
- 1TB Crucial CT1000MX500SSD1 (OS drive)
- 1TB SAMSUNG HD103SJ
- 3TB Seagate ST3000DM001-1ER166

I should state that I have no idea on OC, I am a total newbie and I have already made a disastrous attempt to OC. I had found somewhere someone posting his BIOS settings for the same mobo and cpu as mine in order to achieve 4.2GHz. I just entered the values into my BIOS. Result: Windows booted, then I had several error messages and after a few seconds BSOD. OC killed my SSD OS drive (Crucial MX500).

Now I have a new SSD drive. My question: Is OC safe? If I can *SAFELY* OC my CPU to 3.5-3.7GHz then I would like to try it. But if there is the slightest possibility to kill my SSD again then what CPU would you recommend me to buy? I do not want to spend more than 100euros.

Thank you very much in advance for your help!

 

[EarthDog]

That shouldn't have killed your ssd...

Overclocking isn't without risk, but...its safe overall. Here is a guide for you....https://www.overclockers.com/3-step-guide-overclock-core-i3-i5-i7/

Read it.. then read it again.. then give it a go!

You really need a full upgrade.... so save your 100...overclocking will help, but it isnt a miracle on something that old.

 

[Witchdoctor]

Wow, they are a lot of fun to overclock,

I had that board back in the day, your going hit the 222 BCLK wall ........ it really blows

you can get in the 230's if you really push the PCIe bus like 125 ish ....

Be prepared to deal with a lot of heat , but loving the NB cooler on that board , it was epic

 

[Bill_Bright]

The data stored on the SSD may have become corrupted somehow, but the SSD itself would not have been damaged by what you did. So if truly damaged, something else did it and the timing just made it appear your OCing attempt did it.

As far as your question; "overclock i7-920 or upgrade CPU?", I think only you can answer that. Overclocking your current CPU is free. A new CPU could cost $100s (your current budget limit of ≤ €100 will not get you a significant - if noticeable at all -performance gain).

If me, considering that board has been around for almost 6 years, only supports DDR3, no USB 3.x or M.2, I would not put any money into it. Instead, I might play around with OCing for the learning experience, then save my pennies for a full new upgrade to a new motherboard, new CPU, new DDR4 RAM, new "Gold" PSU (I like EVGA and Seasonic), and a new W10 OS license. You can carry over your current drives and graphics solution until you rebuild the budget to upgrade them too.

 

[trents]

What do you use the system for mostly?

 

[Alaric]

you can get in the 230's if you really push the PCIe bus like 125 ish ....

Every time I've gone over 105 on a PCIe bus my OS got corrupted and I had to reinstall. That was on AMD chipsets, though, so I don't know if that makes a difference.

 

[EarthDog]

A big difference.

 

[Alaric]

OK, I wasn't sure if it was a chipset limitation or a Windows limitation handling the increased speed.

 

[UltraTaco]

Oi oi!! I have 920!! Its in the drawer because I upgraded to xeon x5670 for $17 6 core goodness, but it's an aging dinosaur irregardless..
What's it used for mostly?

You could have run into trouble with overclock because you have all ram sloth occupied. 6 sticks make it hard on integrated memory controller that's located on CPU and perhaps the overclock template you received had also overclock profile for uncore as well(memory controller)

Anything around 3.6-3.8ghz is not very hard though. Mine was able to hold 3.6ghz with default 1.2 voltage. Going to 3.8 needed a bit more, but still decent. You're likely be limited by heat.

 

[wingman99]

Max Turbo Frequency 2.93GHz overclocking to 3.7GHz would give you about 26% performance boost.

 

[UltraTaco]

It only goes to 2.93 on single core *and* when c-states enabled. You usually have it only sit at 2.79ghz, so performance boost will be a little higher, about 32%.

 

[Bill_Bright]

Every time I've gone over 105 on a PCIe bus my OS got corrupted and I had to reinstall. That was on AMD chipsets, though, so I don't know if that makes a difference.

I don't see how it would or could make a difference just because it is an AMD chipset instead of an Intel. The chipset choking on your PCIe setting should NOT corrupt the OS and force a reinstall any more than OCing would permanently damage your SSD. Those settings come into play long before the drive (and thus OS) are touched.

Part of the problem, as I see it, is OCing the PCIe bus affects more than just the PCIe x 16 slot for the graphics card (regardless if AMD or Intel). It affects all devices that use that bus - which includes SATA - depending on the specific motherboard. And since we don't know which AMD motherboard(s) or chipset(s) were referenced here, I don't think it fair to make a blanket statement that suggests corruption of the OS to point a reinstall is needed is a real possibility just because its an AMD chipset, and not Intel.

Of course, I could be wrong. It would not be the first time but I cannot find any supporting evidence one way or the other. So if I am wrong - especially since some of my OCing clients prefer AMD - I sure would like to see some report or white paper or AMD KB documentation supporting the claim that pushing the PCIe bus with AMD chipsets may corrupt the operating system.

 

[EarthDog]

I can't say we have whitepapers, but we can tell you that some platforms versus others had some/a lot/little flexibility when raising PCIe speeds.

This commonly happened to Intel users on that platform when really pushing things. I lived through it all too often...Then Sandybridge came out and, IIRC, tightened down that flexibility (wasn't needed anyway since they were unlocked chips). But yeah, that was the way to increase FSB limitations was to raise the PCIe bus.

 

[Bill_Bright]

but we can tell you that some platforms versus others had some/a lot/little flexibility when raising PCIe speeds.

Right! And that makes total sense. The difference was NOT because it was an AMD platform - as was suggested above.

This commonly happened to Intel users on that platform

Which contradicts the statement made above that claims it made a difference because it was AMD. We now agree, it doesn't. The issue occurs with "some" Intel platforms just as it occurs with "some" AMD platforms.

So what makes a difference is the specific AMD or Intel chipset/motherboard platform - and not because the chipset is AMD.

We are on the same page now.

 

[EarthDog]

So what makes a difference is the specific AMD or Intel chipset/motherboard platform - and not because the chipset is AMD.

let me be more clear...

Some Intel platforms had this pcie flexibility... some didnt (hasnt in generations) AMD never did AFAIK. So, it was platform specific. I could be wrong on the AMD side, but, I don't recall them having that PCIe flexilibty, or the NEED to do so in the first place with their CPUs.

 

[Johan45]

I'm not overly familiar with X58 but I don't think there's one blanket statement here tat only "X" can do it and not "Y" I have OC'd the PCIe bus before to try and get more out of a GPU benchmark on more than one platform. I do believe that the amount of flexibility could be platform related but it's also limited by storage and the card. SOme cards just don't like it other don't care in the slightest. Some platforms like AM4 the BCLK directly afected the PCIe bus requiring a gereation drop for the lanes in BIOS to retain stability and some HW combinations would limit BCLK to just 2 or 3 ticks before they wouldn't boot regardless of dropping to Gen2 or not. Wjile others would allow 150+ on the bclk
Like I said more convoluted than it appears

 

[trents]

My experience with overlocking only goes back about 15 years and is not nearly as extensive or rich as many on this forum. But it seems to me that at different times both AMD and Intel systems have been very overclockable via the FSB/BCLK. To be able to vary the BCLK/FSB more than 2-3 mhz so as to be stable and really overclock a system that way was usually tied to the issue of whether or not the PCI-e bus was locked to the BCLK/FSB. If the two were not in lockstep then it was usually possible to overclock the CPU (even if it had a locked multiplier) considerably via the BCLK/FSB. Having said that, cheaper motherboards typically would not allow raising the BCLK/FSB nearly as much as high end ones.

 

[UltraTaco]

Well, you typically dont need to touch pcie frequency until you push into the limits of overclock. For I never had to change from default 100 when doing g 3.8ghz and now on xeon, she will happily do part 4ghz with default 100 pcie.
Pcie affects hard drives, so it's quite likely raising it caused OS to borke.

 

[trents]

I'm not sure we're talking about the same thing, Taco. If you have a CPU that doesn't have an unlocked frequency multiplier you have no choice but to exceed the BCLK/FSB stock frequency if you want to overclock the CPU.

 

[EarthDog]

Right, but back in those days, there was an option to raise PCIe bus speeds that were separate from the FSB. You can adjust FSB but leave PCIe at 100. When users hit a FSB limit on that platform, raising the PCIe bus helped with that limit and, IIRC, also improved scores a bit in some tests (don't quote me on that, its been 10 years since I touched that platform I think).



Anyway, let's focus on the OP please... whenever he gets back to us.