re: PC Discussion - Gaming, Performance and Enthusiasts

My h220 makes almost no noise. I'm pretty sure the stock fans on the h100 aren't great so on high speed they're pretty loud.

When I had my h80 (pretty sure it uses the same stock fans), it was pretty loud when I put the fan speeds on high for benchmarking.

Is it as loud as the hyper 212?

quote:

---

broke one of my CPUs tonight by dropping it. Bent pins on a FX4100, shitty chip anyway.

---

Good luck. I've had a few people ask me to repair bent pins on various AMD chips over the years, and I was never successful -- not that I really had the proper tools for it anyway. Just my eyes and some small tools of various shapes. When they fall, they fall hard.


HASWELL UPDATE:
At 4.5GHz and a fixed vcore of 1.2v, it is not Prime Stable. But, it did run for about 5 minutes so I saw some potential.

I went back into the bios and changed the voltage from fixed to "Adaptive" (This is something new as an improvement to offset voltage, keeps idle voltage from dropping too low). What it does in synthetic stress tests, though, is add about .09v to my 1.2v. But I just wanted to see what kind of temps and stability I was in for. It's prime stable, and my temps settle at around 68-70. There's a pretty big difference in temps between core 1 and core 4. Core 1 sits at around 75, while Core 4 is about 64. Could be the nature of the chip, poor IHS seating at the factory, or uneven pressure from my Raystorm.

So, 70C at 4.5Ghz at 1.29v (roughly). Not a chip for air coolers.

I think I can do better, though. I need to turn off adaptive vcore and just keep inching it up, along with the multi.

I'm trying to decide what's an acceptable prime temp, knowing that I won't reach those temps in gaming or even folding for that matter. We know that this CPU throttles at 95C and up. We also know that my video card OC (once I redo that) will add a little more heat to the loop even while idle.

I'm not as scared of temps as some water coolers are. It seems like 80C for this chip with a respectable overclock is "normal" -- assuming we aren't talking about insane voltages. As it stands now, I don't think I'll be able to exceed 1.3v without the temps going beyond my comfort zone. With that knowledge, I could just do a fixed 1.3v right now and adjust everything around it.

Ok, now I'm just rambling.

UPDATE:

4.6GHz @ 1.26v. Prime95 blend test, goes through the first set of 10 tests, temps hovering at 60-65 on the hottest core.

Then, the second set of tests starts... within a minute, I see 85, then 90, then 95. I see slight throttling at that point, then BSOD.

Googling more about Haswell stability tests, I read several guides recommending aida64 instead. Never used it before. I downloaded it and ran a stress test. Temps were in the 60 and 70s (but fluctuated A LOT and often). Ran for maybe half an hour, no super huge heat increase like what Prime had caused when kicking it up a notch.

After some more messing around, I got a 4.7GHz stable OC at 1.3v, and AIDA64 temps were fluctuating madly but hovered in the 70s.

So I think I'll stay with this overclock and see how temps look in games, compressing, encoding, and maybe even folding.

Definitely a hot motherfricker. But from what I've read, it appears that I have at least an above average chip. One guide mentioned that the baseline for seeing if you have a "good" Haswell is to set your vcore a 1.25 and see if you can boot to desktop at 4.5GHz. I can, and it's stable.

Funny reddit thread tearing apart an advertisement
You have to be a nerd to appreciate this but I found it funny

For the AMD fan boys

As crazy as it sounds I think you have more control over it by doing that. It is real easy to slip with a razor.

I used a utility blade. Very simple and easy

Here's a nifty comparison.

Just played around with 3dmark11 a little, and used the same GPU clocks as my old setup.

In my old setup:
i7 2600k @ 5.0 @ 1.5v
ASUS P8Z68-V Pro
GPU 1: MSI R7970-2PMD3GD5/OC 3gb @ 1280/1900
GPU 2: MSI R7970 Twin Frozr 3GD5/OC BE 3gb @ 1220/1700
3DMark11 Numbers:
Graphics: 21499
Physics: 10975
Combined: 9511
Score: 16929

New setup:
i7-4770k @ 4.7 @ 1.28v
MSI Z87 MPower
GPU 1: MSI R7970-2PMD3GD5/OC 3gb @ 1280/1900
GPU 2: MSI R7970 Twin Frozr 3GD5/OC BE 3gb @ 1220/1700
3DMark11 Numbers:
Graphics: 21803
Physics: 11695
Combined: 10173
Score: 17526

Graphics increase: 1.41%
Physics increase: 6.6%
Combined increase: 7%
Score increase: 3.5%

Since 3dmark scores can vary by +/- 3%, the increases are mostly negligible. I might try the adobe benchmark later.



Of the 100+ 3dmark11 scores posted with a 4770k and crossfire 7970s, mine is the #1 score by a margin of 153.

Powercolor
Sapphire
Hypothetically what would be the difference in the designs of these cards? Would the Powercolor just dissipate the heat out the back of the case while the Sapphire would into the case?