searching for 5Ghz!!!

[Milkshake86]

So I was doing a little testing and got 4.8Ghz stable around 1.4v over sometime. Temps around 70c but idk just feels too much to me. So i dropped back to 4.5@1.29

[meowN]

how long did you keep it stable for? I could get mine 4hr prime95 4.8 @ 1.38v, and 2hr 5.0ghz @ 1.41v. not worth the jump in voltage/heat to me though.

[Milkshake86]

Only ran prime for 6hrs Im not comfortable w/ 1.4v

[Fierae]

I can get mine working at almost 5, with INSANE voltage, in the end it's just not worth it though.

Now at 4.5Ghz with 1.4v, watercooled though.

[Milkshake86]

Im fairly certain i can bump to 5 @ 1.4 but I mean other than epeen I see no reason yet to hold at 1.4v for long. My biggest surprise was not going over 72c

[meowN]

I can get mine working at almost 5, with INSANE voltage, in the end it's just not worth it though.

Now at 4.5Ghz with 1.4v, watercooled though.

is that for your 3930k or a 2500k? 1.4v seems way too high for that moderate of an OC

Im fairly certain i can bump to 5 @ 1.4 but I mean other than epeen I see no reason yet to hold at 1.4v for long. My biggest surprise was not going over 72c

yeah, as soon as my cpu topped over 80c @ 5.0 I shut it down.

[Milkshake86]

It is not uncommon for some Sandy bridge chips to require a large Vcore bump for 4.5+

Yea I can't go past 58c w/ 4.5ghz only using a corsair A70 to.

[DarkBlade6]

Im running 5Ghz @ 1.48V under load 24/7 , cooled by a H100 in push/pull ... temps are kinda weird but stable , two cores stay under 70C while the other 2 hit 77C after couple hours of Prime95 (ive reseat the waterblock 3 times to make sure it was ... well seated ).

Does it worth it ?! In my opinion Yes, because I only play CPU bound games (MMO Ahah ), in heavy crowded situation or intensive aoe, even at 5ghz, my FPS can drop under 60! While playing Rift we were like 200 ppl (if not more) trying to kill this World event boss in Ember isle and my FPS was around 25... like wtf!! At 4Ghz it would have been Alot lower =/

[Milkshake86]

Whole lot of Vcore there.

[Fierae]

is that for your 3930k or a 2500k? 1.4v seems way too high for that moderate of an OC

That's the 3930 yeah, it takes a whole lot more to get near 5...

Some chips just need a lot of VCore to run higher I think.

[BicycleMafioso]

Generally, Sandy Bridge CPUs are volt-capped, not heat-capped. Nothing strange there.
I also doubt you'd notice any real world difference past 4.5-4.6GHz.

[Deleted]

The highest overclock I've been able to attain, 4.8GHz, runs Prime95 stable for 4 hours at 1.42V. Though, I've never attempted to go above that as I see no reason other than "for fun" (I see no reason to run my 4.8GHz overclock either, which is why I usually boot at either 4.5 or 4.7GHz). My 4.8GHz overclock makes my CPU fairly hot too; it runs somewhere around 75-80c.

The thing is, my 4.5GHz overclock requires a Vcore of 1.22V. It's therefore a ~14% increase to Vcore for a 300MHz overclock which is around 6.2%. However, real performance is barely noticeable; I did a test render (random WoW video) just to see if it would decrease render time in any significant way; it didn't.

[DeltrusDisc]

Im running 5Ghz @ 1.48V under load 24/7

I sincerely hope you're ready to pay for a new CPU sometime. 1.48V 24/7? That's absolutely ridiculous and NOT something I would recommend and I think a lot of other people here, will concur with me on this.

---------- Post added 2012-01-16 at 04:33 PM ----------

The thing is, my 4.5GHz overclock requires a Vcore of 1.22V. It's therefore a ~14% increase to Vcore for a 300MHz overclock which is around 6.2%. However, real performance is barely noticeable; I did a test render (random WoW video) just to see if it would decrease render time in any significant way; it didn't.

The rendering/compressing time? I'm pretty sure that's typically bottle-necked by the HDD write speed, especially when it comes to high OCs on SB chips. :P Correct me if I'm wrong, but that's just what I get from it, like your CPU could probably push the video to render and compress much quicker but ultimately the HDD can only write so fast, but again I might be wrong on this. >_<

[Milkshake86]

Yea I literally sat here dicking around the forums the entire time I was running prime, 1.4v made me nervous as is.

[Deleted]

The rendering/compressing time? I'm pretty sure that's typically bottle-necked by the HDD write speed, especially when it comes to high OCs on SB chips. :P Correct me if I'm wrong, but that's just what I get from it, like your CPU could probably push the video to render and compress much quicker but ultimately the HDD can only write so fast, but again I might be wrong on this. >_<

Doubt it. I render to H.264 .mp4 format at around 7Mbit/s which means that a 8 minute clip (which I believe is what I used when I tested) is around 400Mb. Considering that my HDD can handle Fraps without issues (i.e. ~75Mb/s) it should have no issue to write that 400Mb file in the ~10 minutes it takes to render.

[DeltrusDisc]

I guess it could be the compression process then.

[Deleted]

I guess it could be the compression process then.

Perhaps.

I haven't heard anyone say that increasing clock speed will offer a significant render time decrease to make it worthwhile for the average user (people with a lot better systems than mine).

[DeltrusDisc]

Would RAID0 help cut down the time do you think? Like 2 or 4 1TB 7200RPM drives?

[Milkshake86]

Perhaps.

I haven't heard anyone say that increasing clock speed will offer a significant render time decrease to make it worthwhile for the average user (people with a lot better systems than mine).

Well clearly its because you don't has a magic chip sir! No but back to the topic tbh above 4.3 GHz I havent seen a significant boost in anything realistically. Im sure If I used every tool in the book I could come up with miniscule increase's but from 3.3-4.3 is pretty damn noticable.

[Deleted]

Would RAID0 help cut down the time do you think? Like 2 or 4 1TB 7200RPM drives?

I doubt it. Perhaps Synthaxx can confirm. In either case, I feel we are derailing Milkshake's thread a bit.