Overclocking Preview Bulldozer 8150

[psolord]

Yeah but it means that Bulldozer won't really clock better than SB on conventional air or water cooling.

Which is disappointing, because BD most likely will not have anywhere near the IPC of Sandy Bridge.

You may very well have a point here, but that remains to be seen.

What we should take into consideration here though, is that all this overclocking fanfare, gives hope for high performing/low budget FX 4100 systems, due to their unlocked nature.

At the very least, it could mean competition for Intel! Isn't that great for us?

 

[Edrick]

You realise the last world record, holder, was a Intel Celeron 352 (single core) CPU?

Since then no modern CPU has reached anywhere near that.
Today a new CPU launched, a modern one, with more than 1 core, and managed to beat that record.

I realize that. I guess I am more impressed with OC records with the entire CPU active.

 

[996GT2]

FTFY.

Hybrid water cooling isn't real water cooling. You'll see the same or better result with a high-end air cooler, as we've already seen in the past.

I realize that an Antec Kuhler isn't real water cooling. Let's assume they used the lower-end Kuhler 620. It's still equivalent to a Hyper 212+ and most 2600Ks can reach 4.8 GHz with such a cooler. My 2500K even did 4.8 GHz back when I had a Hyper 212+.

The issue is how a Bulldozer at 4.8 GHz compares to a 2600K at 4.8 GHz.

 

[BlueBlazer]

You realise the last world record, holder, was a Intel Celeron 352 (single core) CPU?

And on LN2 only. Maybe when some overclocker has access to LHe like these guys do (with AMD support), that record will be broken again with Netburst-based single core Celeron.

Since then no modern CPU has reached anywhere near that.
Today a new CPU launched, a modern one, with more than 1 core, and managed to beat that record.

Not 2 "cores", but 1 "module". It does seem that they can't disable individual cores (comes in pairs). Anyway, a single Bulldozer "core" isn't really a true core.

 

[Riek]

I realize that an Antec Kuhler isn't real water cooling. Let's assume they used the lower-end Kuhler 620. It's still equivalent to a Hyper 212+ and most 2600Ks can reach 4.8 GHz with such a cooler. My 2500K even did 4.8 GHz back when I had a Hyper 212+.

i think the issue is mainly user driven here...

One sample doesn't dictate overclocking potential so their is nothing to say about that. The only thing that we can mention is that BD broke a record that hasn't been broken for 10 years orso.

Given this oc wasn't on the final silicon either this unknown oc behavior might even change in a negative or positive way.

The issue is how a Bulldozer at 4.8 GHz compares to a 2600K at 4.8 GHz.

We don't even know the default performance. we don't know how BD overclocks can be tweaked.. What about overclocking with turbo? What about overclocking the l3 cache?

Stop looking for things to be negative about something that hasn't been done in years.

 

[LOL_Wut_Axel]

I realize that an Antec Kuhler isn't real water cooling. Let's assume they used the lower-end Kuhler 620. It's still equivalent to a Hyper 212+ and most 2600Ks can reach 4.8 GHz with such a cooler. My 2500K even did 4.8 GHz back when I had a Hyper 212+.

The issue is how a Bulldozer at 4.8 GHz compares to a 2600K at 4.8 GHz.

True, but...

i think the issue is mainly user driven here...

One sample doesn't dictate overclocking potential so their is nothing to say about that. The only thing that we can mention is that BD broke a record that hasn't been broken for 10 years orso.

Given this oc wasn't on the final silicon either this unknown oc behavior might even change in a negative or positive way.


We don't even know the default performance. we don't know how BD overclocks can be tweaked.. What about overclocking with turbo? What about overclocking the l3 cache?

Stop looking for things to be negative about something that hasn't been done in years.

This. We still need to see if that 4.8GHz is the normal we should expect on moderate voltage.

As far as IPC goes, again, around Llano or a bit lower. You still have twice the cores, though, so it should be faster than Sandy Bridge in multi-threaded, if only by a bit.

 

[996GT2]

True, but...



This. We still need to see if that 4.8GHz is the normal we should expect on moderate voltage.

As far as IPC goes, again, around Llano or a bit lower. You still have twice the cores, though, so it should be faster than Sandy Bridge in multi-threaded, if only by a bit.

When you run 8 threads on a Bulldozer chip, you lose some performance compared to a native 8-core design. Couple that with what will probably be a low IPC and even in multithreaded apps an FX-8150 may not beat a 2600K.

 

[LOL_Wut_Axel]

When you run 8 threads on a Bulldozer chip, you lose some performance compared to a native 8-core design. Couple that with what will probably be a low IPC and even in multithreaded apps an FX-8150 may not beat a 2600K.

Yeah, but according to AMD IIRC, the 200% performance you'd get from two normal cores is 180%. 100% for the first core, 80% when you add the second one. It should still be a tiny bit faster in very multi-threaded stuff.

 

[classy]

Found some more info. It appears over 5G was achieved by all of the cpus they tested using air.

NordicHardware: "It says you reached Frequencies" well above 5 GHz, "with only air and sub-$ 100 water cooling solutions. Mind Going Into more detail about your ventures on air and water?"
Sami Mäkinen : "In terms of CPUZ MHz I've seen virtually every CPU reach over 5GHz when using a good air/watercooler. I believe the highest result I saw with aircooling was around 5.5GHz CPUZ. This CPU hit around 7.8GHz on LN2." Sami Mäkinen: "In terms of CPUZ MHz I've seen VIRTUALLY every CPU reach over 5GHz When Using a good air / water cooler. I believe the highest result I saw with aircooling was around 5.5GHz CPUZ. This CPU hit around 7.8GHz on LN2 '.

http://translate.google.com/transla...ami-maekinen-och-det-nya-vaerldsrekordet.html

 

[LOL_Wut_Axel]

Found some more info. It appears over 5G was achieved by all of the cpus they tested using air.



http://translate.google.com/translate?hl=sv&sl=auto&tl=en&u=http%3A%2F%2Fwww.nordichardware.se%2Ftest-lab-cpu-chipset%2F44128-intervju-med-amds-sami-maekinen-och-det-nya-vaerldsrekordet.html

5.5GHz is the highest on air cooling, then? That's probably at very high voltage, and with a chip which would degrade rapidly and would probably not be stable for 24/7 use. Given that, I think around 4.8-5GHz is what you'd want to do on moderate voltage for 24/7. It's 300-500MHz higher than SB, but like I said earlier, you'd need at least 6GHz to get close to Sandy Bridge on single-threaded. To beat it you'd need 6.5GHz if we're to assume Thuban IPC. :\

 

[AtenRa]

For all we know the IPC could be worse than Phenom II

yes, and then we wake up

 

[Arkadrel]

For all we know the IPC could be worse than Phenom II

JF-AMD claims its higher (bulldozer IPC > Phenom II).
Mhz for mhz, bulldozer is supposed to be faster.

How many Phenom II 's reach 5ghz with overclocking?

Also just imagineing a Phenom II @5ghz... that in itself wouldnt be a bad cpu.

 

[classy]

5.5GHz is the highest on air cooling, then? That's probably at very high voltage, and with a chip which would degrade rapidly and would probably not be stable for 24/7 use. Given that, I think around 4.8-5GHz is what you'd want to do on moderate voltage for 24/7. It's 300-500MHz higher than SB, but like I said earlier, you'd need at least 6GHz to get close to Sandy Bridge on single-threaded. To beat it you'd need 6.5GHz if we're to assume Thuban IPC. :\

Many of you are not true overclocking folks. So far this appears to be a rather hum hum cpu, but when I read this, this is no matter how you slice it, is as impressive as hell. I remember many many nights going to work tired as hell trying to push a cpu as far as I could. I reset a bios so many times I developed near carpal tunnel type symptoms, lol. This is an impressive feat. This is about the overclocking potential of this cpu not its performance. And until we have some true numbers all these posts about performance are guesses at best. And most of the SBs don't do 5G without some serious cooling themselves. And whether or not any of them Intel or AMD overclocked that high would be truly 24/7 reliable also would be a shot in the dark.

 

[busydude]

They disabled 3 modules to hit 8.4 GHz.

So it wasn't an 8-core hitting that clock, but rather a dual-core.

Lame.

Well, that is how overclocking is done.

 

[LOL_Wut_Axel]

JF-AMD claims its higher (bulldozer IPC > Phenom II).
Mhz for mhz, bulldozer is supposed to be faster.

How many Phenom II 's reach 5ghz with overclocking?

Also just imagineing a Phenom II @5ghz... that in itself wouldnt be a bad cpu.

Yes, but even if it has 5% higher IPC than Phenom II (that's the most that's realistic at this point), you'd get 180% performance from a module in comparison to two Phenom II cores. That means you'd have a 15% performance deficit with two cores in comparison to Phenom II. That can be mitigated with higher clock speeds, and comparing a BD at 5GHz with a PII at 4GHz would mean it'd be 10% faster comparing a module to two cores (15% deficit by default, but 25% higher clock speed=10% higher performance).

Many of you are not true overclocking folks. So far this appears to be a rather hum hum cpu, but when I read this, this is no matter how you slice it, is as impressive as hell. I remember many many nights going to work tired as hell trying to push a cpu as far as I could. I reset a bios so many times I developed near carpal tunnel type symptoms, lol. This is an impressive feat. This is about the overclocking potential of this cpu not its performance. And until we have some true numbers all these posts about performance are guesses at best. And most of the SBs don't do 5G without some serious cooling themselves. And whether or not any of them Intel or AMD overclocked that high would be truly 24/7 reliable also would be a shot in the dark.

If me not liking highly degrading chips in a short time means I'm not a true overclocking folk, then by all means, I'm not. If I get an SB CPU I'm sure as hell not gonna run anything higher than 1.35V through it. I want a CPU that will last me with no problems for at least five years, since I'm not gonna just throw it out when I want to get something new. I don't like selling my old stuff, so yeah.

Still, I think 4.8-5GHz on moderate voltage for Bulldozer is about right.

 

[guskline]

Many of you are not true overclocking folks. So far this appears to be a rather hum hum cpu, but when I read this, this is no matter how you slice it, is as impressive as hell. I remember many many nights going to work tired as hell trying to push a cpu as far as I could. I reset a bios so many times I developed near carpal tunnel type symptoms, lol. This is an impressive feat. This is about the overclocking potential of this cpu not its performance. And until we have some true numbers all these posts about performance are guesses at best. And most of the SBs don't do 5G without some serious cooling themselves. And whether or not any of them Intel or AMD overclocked that high would be truly 24/7 reliable also would be a shot in the dark.

I agree classy. I have a SB 2500k running at 4400 all the time. I have subjected it to every torture test known and it has passed with flying colors. I'm sure with serious tweaking I could go higher ( I booted into Windows at 4.8 but it froze on some tests). The point is that perhaps AMD now has a CPU with serious overclocking potential. My Phenom II 965s are only overclocked by 300 (3.7 vs 3.4) while my SB is overclocked by 1100(3.3 vs 4.4). The Phenom IIs don't have much overclock potential though they have unlocked multipliers. The BD sounds like it has much more headroom. This is beginning to sound like serious competition to the SB!

 

[Ferzerp]

I agree classy. I have a SB 2500k running at 4400 all the time. I have subjected it to every torture test known and it has passed with flying colors. I'm sure with serious tweaking I could go higher ( I booted into Windows at 4.8 but it froze on some tests). The point is that perhaps AMD now has a CPU with serious overclocking potential. My Phenom II 965s are only overclocked by 300 (3.7 vs 3.4) while my SB is overclocked by 1100(3.3 vs 4.4). The Phenom IIs don't have much overclock potential though they have unlocked multipliers. The BD sounds like it has much more headroom. This is beginning to sound like serious competition to the SB!

Does that mean that netburst celerons are serious competition to SB? Cherry-picked ones could OC to similar clock speeds.....

 

[BlueBlazer]

I think you are all forgetting something. They could be overclocking with only 2 "cores" (1 "module") enabled all the time (whether on air, water, LN2, LHe). :hmm:

 

[guskline]

Ferzerp, of course not. The Celeron is a much older design. Don't get me wrong, I love my SB 2500k and will be surprised if the BD beats it but right now SB stomps the Phenoms on most diagnostics. I think this OC demo might be showing that the BD has significantly closed the gap. Only time and the release of the BD for objective testing will tell.

 

[dma0991]

I wonder if this would push Intel to release a fully unlocked SB to push beyond >9000MHz?

 

[guskline]

I wonder if this would push Intel to release a fully unlocked SB to push beyond >9000MHz?

What is the record for SB overclocking?

 

[LOL_Wut_Axel]

I think you are all forgetting something. They could be overclocking with only 2 "cores" (1 "module") enabled all the time (whether on air, water, LN2, LHe). :hmm:

I sure hope not.

 

[sawtx]

I think you are all forgetting something. They could be overclocking with only 2 "cores" (1 "module") enabled all the time (whether on air, water, LN2, LHe). :hmm:

They were running "fully multi-threaded" benches at ~5GHz on air, which I would assume equals all cores were active.

 

[Kaekae]

We want IVYBRIDGE! That overclocking feat is Impressive though.....

 

[garagisti]

Before Bulldozer set this world record, we had a whole load of wankers claiming it may not overclock as well as SB. Now, we have IPC in question. Anything else (including the kitchen sink) you want to throw at AMD? If only people bothered to read, there were 8 processors which did more than 8 Ghz in the lot provided to those lads. Almost 100% rate of 4.5-5 Ghz on air is pretty good. There were also a fair few that reached 7+ Ghz.

I guess the sour pusses here are just going to find something or the other to find fault with... You want to buy Intel SB/ SB-E/ IB, then go those threads and rave about them. You have something in particular to rant about AMD's processors, by all means do so in respective threads. However, ranting in a "world record" thread about how you're not getting a happy ending is just plain douche-baggery on your part.