Intel Core i5-750 Overclocking Preview @ [H]

[FrgMstr]

Intel Core i5-750 Overclocking Preview - Our first preview of our Intel Core i5 processor that we purchased online last week. We put the 2.66GHz Core i5 model 750 through some overclocking paces to show you what can be achieved easily with an enthusiast air cooler.

I have very much impressed with our first overclocking experience with a retail version of Intel’s Core i5-750. Keep in mind that we wanted to show you air cooled results here but did not have the proper mounting brackets for our Thermalright TRUE cooler. This lack of pressure on our mating surfaces surely could have surely impacted our results. I do however have no problem suggesting that with a little BIOS tweaking and maturation, and good cooling, that many of you will be seeing fully 100% load stability at 4.2GHz and possibly even 4.4GHz on air with the Intel Core i5-750.

 

[magoo]

Slick.

Typo fixed, thanks for the extra eyes. - Kyle

This CPU and MB combination looks very excellent. Can't wait to see the comparison to the Core i7.

 

[Compddd]

Shouldn't i5 be slower clock for clock compared to i7?

 

[Riccochet]

Cool beans

 

[Deleted member 82943]

Kyle,

awesome write up as always. I think the one-button thing is really neat and if stable I figure OEMs and boutique builders might offer it as well.

did you run any benches? I ask because I know you probably can't show us the numbers but can you say *subjectivly* that it is faster or not than the a comparable chip say an i7 920?

Alex

 

[FrgMstr]

Shouldn't i5 be slower clock for clock compared to i7?

Should it be?

Both Nehalem microarchitecture.

i7 = 4C/8T i5 = 4C/4T

i7 = Triple Channel DDR3 i5 = Dual Channel DDR3

So......unless you get into applications that are thread dependent or memory bandwidth dependent....who knows.

Hopefully up tomorrow. But it will be without metrics.

 

[FrgMstr]

did you run any benches? I ask because I know you probably can't show us the numbers but can you say *subjectivly* that it is faster or not than the a comparable chip say an i7 920?

Yep, our usual mobo benchmarks. Graphs up tomorrow, clock for clock, mem for mem, no metrics, but we will show you bar graphs.

 

[Sphere]

Looks good, but 1.4v for 3.6 is pretty damn high, those temps must have been 70-80'ish under load.

 

[niconx]

Safe
Not safe
Really not safe
but no Really not safe oh crap call the fire department?

 

[FrgMstr]

Looks good, but 1.4v for 3.6 is pretty damn high, those temps must have been 70-80'ish under load.

I do think that at 3.6GHz could have been done at stock voltages......just like I typed in the article. RTFA baby!

 

[Sphere]

That's a different story,50-60 ish isn't too bad, considering you were using air.

 

[Deleted member 82943]

how much did your new baby set you back?

 

[BlackTTC6]

Very nice

 

[XvCluTchvX]

nice, but those volts r too high

 

[Fafeifa]

OC sucks compared to D0 i7 specially considering there is no HT

 

[xrealm20]

still interesting to see what happens with this new chip -- if the pricepoint is good enough, it may be worth the upgrade for people who are still running the core 2 series. Obviously if you already have an i7, this product isn't for you

 

[Sojuuk]

Was pretty impressed. Unless the 860 is somehow magical I wont be buying it.

OC sucks compared to D0 i7 specially considering there is no HT

take the mounting bracket off the i7 920s and tell us how that goes.

 

[ThisIsTheNewSound]

Nice writeup. Looks like this thing is going to be an overclocking champ and an amazing value for the price.(once it stabilizes, of course) Can't wait to get my hands on one.

 

[Kendrak]

Can you please for everything that is holy benchmark what kind of ppd you get running notfreds in a VM?

We know what an i7 will do with HT, but i5 is a mystery. Your folding teams upgrades hang in the balance.

 

[vengence]

This lack of pressure on our mating surfaces surely could have surely impacted our results.

I'm thinking you only want 1 surely in that sentence. Or for that matter, "This lack of pressure on our mating surfaces could have impacted our results." is clean, efficient, and gets the point across as well

Beyond that was hoping for better after seeing the 920 D0s.

However, with the integrated PCI-E controller when can we expect to see benchies seeing if this has any affect on gaming in both single and multi-GPU configurations. i.e 920 vs 750 both at same blck same multiplier and also an i5 at say 12x200 and 20x120 in both single GPU and sli/crossfire settings? I don't think you would need to do the "real world" benchmarks as a canned time demo could give you meaningful results, or at least show where you need to spend time looking at real world gaming.

 

[xxjudgmentxx]

Nice. Just waiting for AMD to drop their prices again

 

[TGA]

Can you please for everything that is holy benchmark what kind of ppd you get running notfreds in a VM?

We know what an i7 will do with HT, but i5 is a mystery. Your folding teams upgrades hang in the balance.

I am quite sure you wont see any numbers until NDA is up.

However unless I am overlooking something clock for clock the i5 750 with it's 4 cores should get very similar results to an i7 with HT disabled in your situation.

 

[FrgMstr]

That's a different story,50-60 ish isn't too bad, considering you were using air.

I have the ASUS P7P55D Deluxe running incubated (at 50C) for 48 hours now and under full CPU load here are the temps at stock clocks. This is with a Koolance Exos2 and Koolance 345 waterblock.

how much did your new baby set you back?

The Core i5-750 was $205 at Fry's.

 

[§kynet]

This lack of pressure on our mating surfaces surely could have surely impacted our results.

Are you SURE?

Looks good, but 1.4v for 3.6 is pretty damn high, those temps must have been 70-80'ish under load.

Seeing temps like that are downright scary to me. But building ~80% AMD systems myself, I guess I'm not used to seeing anything near those kinds of temps.

I'm curious to see 100% stable overclocks and associated heat output.

 

[xrealm20]

Are you SURE?


Seeing temps like that are downright scary to me. But building ~80% AMD systems myself, I guess I'm not used to seeing anything near those kinds of temps.

I'm curious to see 100% stable overclocks and associated heat output.

70-80 on the i7's isn't too out of the ordinary. I'm guessing that with a proper mount, the temps should be around 50-60 Celsius

 

[§kynet]

70-80 on the i7's isn't too out of the ordinary.

Yep. It's just my brain is programmed to see anything over 60 as hit the panic button. Kinda like GPU temps, it took me a bit to see 80+ load temperatures as normal.

I don't know about a drop of 20 degrees though, that would have to be some poor contact, more like floating.

 

[Dynafrom]

aww, I just bought a i7 920

are the new chips a stepup from i7?

 

[FrgMstr]

Are you SURE?

Stop calling me Shirley!

 

[FrgMstr]

aww, I just bought a i7 920

are the new chips a stepup from i7?

There will be no reason in my mind for anyone to move from an i7 to an i5. And if you bargain hunt, you can build an i7-920 rig for very close to what an i5-750 is going to cost you at launch.

These new P55 motherboards are looking to be some real ass kickers though. I think 250MHz BCLK rates are going to be within reach on many of them. Don't quote me on that though, it is just a gut feeling.

 

[FrgMstr]

250BCLK would be a big deal. I'm not sure how to balance that vs the tripple channel memory you get with an i7 though....

To be perfectly honest with you, I have found no reason as to why any desktop user needs all the bandwidth afforded by an i7 system. Even at a measly DDR3 speed of 1066MHz, the i5-750 system delivers 13.7GB/s and 20GB/s when overclocked to 3.2GHz-CPU and 1600MHz memory.

The footprint easily afforded by the i7 might be arguable, but I have only been able to pull more than 8GB one time in real world usage scenarios.

That all said, Gigabyte's high end P55 board has six memory slots feeding the dual channels on the P55.

 

[Exeodus]

Weren't some of the Lynnfields supposed to support Hyperthreading?

 

[CaptNumbNutz]

Kyle-

Your using a TRUE Copper without proper mounting bracket (unavailable atm?), and the fan was mounted to the side...

Is it possible, considering the HS weight, to use zip ties through the motherboard holes to achieve "almost enough" contact pressure? This could be an acceptible alternative for the time being given the fact you guys bench with the motherboard in the horizontal position. At the very least this might secure the HS in one spot and allow you to mount the fan on the heatsink itself without worrying about the fan vibrating it out of place.

 

[vengence]

This is true. I've often thought that the Core i7 was overkill for general desktop usage and gaming. (Nothing wrong with that mind you.) However I've always thought that the server and high end workstation market could certainly benefit from what Nehalem brought to the table. Its advantages are certainly apparent for running virtual machines and databases. For the occasional word document, web browsing and gaming, that advantage is virtually zero over the Core i5.

I am curious however, if the Core i5 would provide the same or similar results as an overclocked Core i7 does when running multi-GPU setups at maximum settings. The Core i7 certainly allows for multi-GPU configurations to scale better than the Core 2 Quad's did, so I wonder if the Core i5 can do the same thing. I'd assume so, but to what extent it can do this compared to the i7, I do not know.

I wouldn't recommend anyone looking for an internet box + casual gaming to bother with an i7. That doesn't make any sense. Frankly, if your doing light office usage and mild to medium gaming, anything over a stock E8500 dual core is probably overkill.

I think there are is a small, but significant portion of the community aka "enthusiasts" who do things like VMs, solve computationally intensive math problems, and hardcore gaming (3 very different problems).

I've seen a lot of evidence lately that some games are being CPU limited by the "old" quads at "only" 4.0Ghz. Frankly, I don't have enough time or money to be able to figure out exactly if an i7 is faster in some games with multi-GPU setups than a similar quad because of the integrated memory controller, faster clock per clock CPU power, more cores, or because it has more penguins. With the integrated PCI-E controller on the i5 I know things are going to get even more complicated.

 

[FrgMstr]

Kyle-

Your using a TRUE Copper without proper mounting bracket (unavailable atm?), and the fan was mounted to the side...

Is it possible, considering the HS weight, to use zip ties through the motherboard holes to achieve "almost enough" contact pressure? This could be an acceptible alternative for the time being given the fact you guys bench with the motherboard in the horizontal position. At the very least this might secure the HS in one spot and allow you to mount the fan on the heatsink itself without worrying about the fan vibrating it out of place.

Yes that is possible. This is what we have done with all all water cooling for LGA 1156. The size of the base of the TRUE and the proximity of the pass-through holes make it where I don't think it makes much difference in mating.

I've seen a lot of evidence lately that some games are being CPU limited by the "old" quads at "only" 4.0Ghz.

Please provides links to the "lot of evidence" that you have seen on this topic. I would like to read up on this myself.

Our last article on this is here.

 

[Impulse]

aww, I just bought a i7 920. are the new chips a stepup from i7?

No, they're a (mild) step down, the only reason to go with an i5 is the cost of the processor itself if you opt for an i5 750 or a dual core i5 (otherwise the i7 920 can be had for the same price as higher end i5 stuff) or if you want to save money on the mobo (since P55 mobos should, eventually, be $50-80 cheaper than X58 mobos).

Weren't some of the Lynnfields supposed to support Hyperthreading?

From what I understand, yes, just not the one Kyle got at Fry's... The 750 was always intended to be the cheapest quad core i5 (at $200-ish), Intel disables HT on it to differentiate it a bit more from the chips higher up the price scale. The next quad core after it ends up costing the same as a 920, so it might be a way for Intel to not undercut it's own X58 platform as much too.

 

[Unoid]

looks like my q6600 @ 3.6ghz will continue to last

 

[Jthm27]

Next week is the day =D

 

[Sojuuk]

Next week is the day =D

that doesn't quite make sense.

 

[RickyJ]

that doesn't quite make sense.

NDA lifts next week, not sure what day. That's what he was getting at!

Looking forward to more results, Kyle! Definitely looks like I'll be getting an i5 soon, I use i7's and C2Q's at work and my poor old AthlonXP isn't cutting it at home anymore!

 

[Impulse]

Where can you get an i7 920 for $220? I know Microcenter had 'em for $200 or w/e for a while but a good chunk of us don't live anywhere near a MC. The 750 is the only quad i5 that makes any sense for the average enthusiast, because if you opt for anything higher you're paying the same as you would for a retail priced i7 920 and at that point paying $50 more for an X58 mobo doesn't amount to much. P55 mobos are probably gonna carry a premium for the first 3-6 months anyway so I'll be surprised if they're much more than $30 cheaper than the cheapest X58 boards out now.

P.S. I'll gladly donate a 75GB 75GXP Deathstar to anyone that wants to destroy it in a creative way! Last HD that ever died on me personally, except for my old S2 TiVo's drive I guess.