Lynnfield benchmarks up

[jennyh]

doesn't the Nehalem have the cold bug still??
not being able to reach critically low temps would not allow it to OC so high

I wasnt aware nehalem ever had the cold bug, but somebody else will be able to confirm that one for sure.

 

[fazers_on_stun]

Lol what?

You have links to back that up ofc.

No, but I'm sure if you google or bing it, you'll quickly find the November 2008 i7 review links 😀.

Actually I do have a bunch of links on my work computer, but it's 10PM here in Wash. DC and even us wage slaves are home relaxing from work . So tomorrow maybe.

 

[TechnologyCoordinator]

Just want to add, you can now buy water cooling that is hardly different from standard air cooling.

It would still have tubes with water (actually there are additives to prevent corrosion). There is just more maintenance and worries when you have water tubes running through a computer, but maybe I'm just too faint of heart! I was always a weak overclocker, always playing it safe (still managed to fry my Socket A board though).

That's just my personal preference though, not saying all people would have my fear of water cooling.

 

[jennyh]

but to keep LN2 going, you would need a compressor to make N2 into a liquid, not really feasible in the common house (let alone in a computer case)

And 10 years ago if you told me i could have cyclone type vacuum cleaning in my hands i'd have laughed loud. What about sizzling stuff to a cooked state in a microwave 20 years+ ago? Before then it would have been like magic.

Progress happens, everywhere.

 

[TechnologyCoordinator]

TC - It's not just about binning, it's about improving too. Intel is more about binning than AMD is, AMD tend to refine a lot during a process whereas intel tend to do it in splits, like today.

I don't agree. Do you have anything to back up your statement that AMD refines during a process?

I think AMD does more improving with their Tick-Tock strategy. You got a whole new arch (Tick), then you get a die shrink and new features (tock), and then the process starts over again.

I think AMD is coming out with things like the 965BE in the same way that Intel "innovated" with Prescott, they are just raising the ceiling of power and heat.

The last major improvement I've seen from AMD was the PII (which are good processors).

 

[mindless728]

I wasnt aware nehalem ever had the cold bug, but somebody else will be able to confirm that one for sure.

Nehalem at -120C

couldn't find anyone using LHe like for AMD, though looks like Nehalem can use LN2, and there is a 72C difference between LN2 and LHe (LHe being lower)

 

[TechnologyCoordinator]

And 10 years ago if you told me i could have cyclone type vacuum cleaning in my hands i'd have laughed loud. What about sizzling stuff to a cooked state in a microwave 20 years+ ago? Before then it would have been like magic.

Progress happens, everywhere.

I think we'll see air cooling get further refined, but I don't think we'll ever seen mainstream computing move beyond air cooling. Like I said, I think it would be more likely to see the opposite, to see low-end mainstream computers (the ones with IGPs and such) move to passive cooling to eliminate moving parts and decrease failures.

 

[The_Blood_Raven]

Ok here is the truth.

i5 is a superior platform than dragon at that price range because it is FAR faster clock for clock (2.66 vs 3.4 Ghz)

Wut? At 2.66ghz, the i5 loses every single benchmark vs the 3.4ghz PII. It is only when the i5 is *overclocked* that is catches up. Turbo = overclocking, there is NO difference at all.

In multi threaded apps, the processor is running at 2.66 Ghz with 4 cores, which is what I was refering to. In games, it is running at 2.9 Ghz with 3 cores active (2 for games plus another for background) or 3.0 Ghz (assuming that the entire system plus the game is being handled by 2 cores) and you still get a good amount of clock for clock superiority.

it overclocks higher (hitting 4.0 Ghz more consistently)

You know that after 1 days benchmarks? I very much doubt the i5 750 is as good an oc'er as the 965 BE tbh. Normally I go with Anand and theirs got to 3.9ghz. The first 940 BE on Anand got to 4ghz, it was only when they shifted all their benchmarks to 64-bit the overclocks started to suffer.

I've overclocked a lot of Phenoms, getting over 3.8 is rare and requires a lot more voltage which means that the overclocks are limited by more factors. You are right I can not give you a 100% sure answer here, but from what I have seen it does a bit better.

 

[fazers_on_stun]

My take on Turbo mode is that it IS a feature. However, it also IS overclocking. Overclocking is taking the clock on the box and going beyond it, imo. However, Turbo Mode is an overclocking feature that is insured by Intel which is why EVERY core i5-750 can reach:

3.2 with one core in use
3.06 with two cores in use
2.93 with all cores in use

I don't know if you all will think of this as a similar example, but I kind of think about it in terms of speedstep and Cool and Quiet. If we were arguing about wattage instead of CPU superiority, and only one of the CPU's had underclocking & undervolting abilities, people would be bound to claim that the comparison was unfair. In a similar way, I think of Turbo as reaching THE HIGHEST THRESHOLD of performance with 1-4 threads WITHIN the defined TDP.

Excellent point, dna708! Turbo is the same as speedstep, only in the opposite direction.

And yes AMD wishes they had thought of it first 😀. So now they have to copy Intel on yet one more item... 😗

 

[jennyh]

I don't agree. Do you have anything to back up your statement that AMD refines during a process?

I think AMD does more improving with their Tick-Tock strategy. You got a whole new arch (Tick), then you get a die shrink and new features (tock), and then the process starts over again.

I think AMD is coming out with things like the 965BE in the same way that Intel "innovated" with Prescott, they are just raising the ceiling of power and heat.

The last major improvement I've seen from AMD was the PII (which are good processors).

You are probably right, but mby not too...

I mean we had a ddr2 940, ddr3 945, ddr3 955, ddr3 965 @ 140w tdp. The memory controller at least changed right from the start, and the TDP has changed on the 965.

Compare that to ATI's graphics, and you get something more like what intel do. 4870 > 4890 with about 9-10 months gap. Thats the same gpu with one big revision, much like i5 compared to i7.

 

[jennyh]

Excellent point, dna708! Turbo is the same as speedstep, only in the opposite direction.

And yes AMD wishes they had thought of it first 😀. So now they have to copy Intel on yet one more item... 😗

Uhh..you gotta be kidding me fazers. Ever heard of cool n quiet?

 

[jennyh]

Gawd it's late, nn and hf fanbois ;P

 

[mindless728]

Uhh..you gotta be kidding me fazers. Ever heard of cool n quiet?

my guess is he was talking about Turbo Boost as something AMD wish it thought of first

 

[fazers_on_stun]

You don't go over TDP while overclocking...Good grief.

Please will you just actually read up on what TDP actually means instead of believing what some forum moron once told you.

TDP = the *maximum*, that includes insane overclocks. Silly little day-to-day overclocks don't even get close. If you make me drag up a bunch of reports to prove this i will be *incredibly* pissed and you will never, ever hear the end of it again.

IIRC, "TDP" = Thermal Design Power, which is the max power the CPU can safely dissipate with the stock cooler within safe temperatures. That's why the 95 watt TDP i5 comes with a tiny, dinky little cooler whereas the 130 watt TDP i7-920 et al come with relatively huge coolers if you buy the retail package.

If you use water, phase-change or LN2, you can safely dissipate much larger amounts of heat/power without exceeding the safe temperature range. So yes overclocking on good air, water, phase or LN2 will exceed the TDP yet temps remain safe.

Now electromigration and other consequences of exceeding the max VCore is another story 😀.

 

[TechnologyCoordinator]

I mean we had a ddr2 940, ddr3 945, ddr3 955, ddr3 965 @ 140w tdp. The memory controller at least changed right from the start, and the TDP has changed on the 965.

AMD in many respects [strike]have[/strike] has been much better about upgrade paths.

I generally buy a motherboard with this in mind:

- I'll upgrade the processor once
- I'll upgrade the video card once
- I'll throw in more RAM at some point


I recently bought a 775 motherboard with a Q8200 processor. I think I probably won't be able to upgrade my processor (especially now in light of what i5 is like). So I got burned this time because I bought into a socket that was near EOL. It's partially my fault, but due to a hardware failure I had to get into a new motherboard and back in March I had to choose from 775 and AM3.


(Edit: Hey Jenny, I committed my pet peeve above and had to correct it! 😀 )

 

[mindless728]

AMD in many respects [strike]have[/strike] has been much better about upgrade paths.

I generally buy a motherboard with this in mind:

- I'll upgrade the processor once
- I'll upgrade the video card once
- I'll throw in more RAM at some point


I recently bought a 775 motherboard with a Q8200 processor. I think I probably won't be able to upgrade my processor (especially now in light of what i5 is like). So I got burned this time because I bought into a socket that was near EOL. It's partially my fault, but due to a hardware failure I had to get into a new motherboard and back in March I had to choose from 775 and AM3.


(Edit: Hey Jenny, I committed my pet peeve above and had to correct it! 😀 )

for me, its a "i might" throw in another (or much better different) video card
i usually OC the crap out of a cpu (this time PII X4 810)
i have plenty of ram from the get go (8GB, last build was 6GB)

 

[TechnologyCoordinator]

i usually OC the crap out of a cpu (this time PII X4 810)

I'm a wuss, I killed one MB OCing, and noticed short life on a couple others that I did moderate OCs on. But this may be in part to my lack of OCing skill. So I'm afraid of OCing because I value the motherboard life more than the extra performance. Although once I decide I need to upgrade I do play around since I'm about to replace it anyway.

 

[The_Blood_Raven]

http://www.amdzone.com/phpbb3/viewtopic.php?f=52&t=136697&st=0&sk=t&sd=a

Good god the horror:

"All review sites except lostcircuits are paid off by Intel and are biased, our scientists prove it"

We need to develop a nuclear strike that totally destroys a website, because AMDzone is an insanity ward. Remember when they were actually... SANE? AMDzone used to be a decent website with a few glaring exceptions.

 

[mindless728]

I'm a wuss, I killed one MB OCing, and noticed short life on a couple others that I did moderate OCs on. But this may be in part to my lack of OCing skill. So I'm afraid of OCing because I value the motherboard life more than the extra performance. Although once I decide I need to upgrade I do play around since I'm about to replace it anyway.

i have an ASUS A8N-SLI Premium that i used to OC the 4000+ from 2.4GHz to 3.2GHz, and i kept it like that for about a year, still kicking along as my file server

though soon to be replaced by an athlon II x2 240 that is going to be a HTPC/File server hybrid monstrosity

http://www.amdzone.com/phpbb3/viewtopic.php?f=52&t=136697&st=0&sk=t&sd=a

Good god the horror:

"All review sites except lostcircuits are paid off by Intel and are biased, our scientists prove it"

We need to develop a nuclear strike that totally destroys a website, because AMDzone is an insanity ward. Remember when they were actually... SANE? AMDzone used to be a decent website with a few glaring exceptions.

went there today, couldn't stand it, and i call myself an "AMD Person"

 

[The_Blood_Raven]

I'm a wuss, I killed one MB OCing, and noticed short life on a couple others that I did moderate OCs on. But this may be in part to my lack of OCing skill. So I'm afraid of OCing because I value the motherboard life more than the extra performance. Although once I decide I need to upgrade I do play around since I'm about to replace it anyway.

Its all about the cooling, I have upgraded a lot of northbridge coolers which is a practice that is far too ignored by the overclocking community. Also if you ever had an nforce board (yes even the 700i series) then any form of use, even at stock speeds, kills the damn things. I absolutely HATE nVidia boards, but I digress.

 

[mindless728]

Its all about the cooling, I have upgraded a lot of northbridge coolers which is a practice that is far too ignored by the overclocking community. Also if you ever had an nforce board (yes even the 700i series) then any form of use, even at stock speeds, kills the damn things. I absolutely HATE nVidia boards, but I digress.

hey old school NForce4 on S939 were pretty good (my server is one)

a previous build i had (sold to a friend) had NForce5 series that wasn't too bad

 

[TechnologyCoordinator]

i have an ASUS A8N-SLI Premium

LOL, that's the same one I had. For a while I took my Opty 175 from 2.2 to 2.45 with only a minimal voltage increase on stock air.

 

[fazers_on_stun]

However, intel get to post lower idle wattages, and bribed benchmarkers can post benches with the cpu described as 'i5 750 @ 2.66ghz' when in actual fact it is operating at 3.2ghz. Since they got bunged a huge wad of cash, that's all ok. ;p

Just remembed that when you buy intel, a certain percentage of what you paid goes on their legal fees, bribery and global warming.

LOL - now you made me laugh!

Next time I'm not in Ireland, remind me to buy you a green beer!

I've lived in London, Munich and Cadiz, besides too many places here in the US, but never in Ireland!

 

[mindless728]

LOL, that's the same one I had. For a while I took my Opty 175 from 2.2 to 2.45 with only a minimal voltage increase on stock air.

the board i have now is pretty decent (M4A78-E, DDR3 was a little too expensive when i got this), i can go from 2.6GHz to 3.2GHz without a voltage increase (from stock 1.325)
though at stock 2.6GHz i use 1.2v

 

[someguy7]

I was simply making a point.

You don't actually believe that the 965 BE is somehow worse than a 955 BE, or 940 BE do you? Just because it has a 'TDP' of 140w instead of 125w? All that means is, this cpu is capable of drawing more watts under exceptional circumstances than any other Phenom II could. The reason why it can do that is because it's capable of reaching higher overclocks.

The 965 is a binned from the same wafers is the 955 and other PII's. Its basically a factory overclocked 955. Take the 965 and clock it down to a 955. It will no longer draw 140w. The reason the 965 or any top binned cpu can usually overclock higher is simply because it is the best silicon from the batch. Time+process improvements will give you 965BE 125w version. Don't be surprised if a 975BE 140w then comes out.

None of these exotic cooling methods will be mainstream. Extreme overclockers been using them for years. They will never be in consumer machines. They are used to max overclock runs. Thats it. The cpu itself would not last with the extreme volts needed for such runs either.