• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Bulldozer Finally!

Bleek said:
Could it not simply be the fact that the 8 core will be 'denser' therefore negating the process transition/drop?
Click to expand...

But they have a 95W 8 core part :p?

My problem is the utter pointlessness of it.
It consuming more power and being hotter would do nothing for its OC'ing ability except hamper it.

Case in point, the 95W 1055T's versus the 125W ones.
 
mjones1 said:
I dont think it's official but there is hard evidence from many motherboard manufactures that the FX-8120 @ 3100MHz is 95w.
Click to expand...

That's my point.
I know the 8150 (If it's real) and 8170 (If its real) is 125W, which note I didn't bring into the equation, due to its core speed, I'd expect nothing less, and it's still equally as impressive.


My point was only with this CPU.


CAT-THE-FIFTH said:
The thing is that the 4 core Bulldozer is going to have the best IPC as it has twice the L3 cache per core than than the 8 core version.
Click to expand...
I've already said the most interesting comparison if the BD 4 core and the i5 2500k quite a while ago. At least, for me it is, as it's a true "4 v 4, best architecture wins" SB could have an IPC advantage, but BD could have the clocking advantage.
 
CAT-THE-FIFTH said:
The thing is that the 4 core Bulldozer is going to have the best IPC as it has twice the L3 cache per core than than the 8 core version.
Click to expand...

Do you know if this is 'fixed' cache per core though ? it maybe that if its using 4 cores it will drop the cache addressing to the remaining cores or keep it for example at 1kb and use the rest of the cache for 4 cores only.....seems plausible
 
Bleek said:
Isn't there a difference in how AMD and Intel measure TDP anyway, I was led to believe Intel are more 'forgiving' with their calculations?

In reality the 125W Phenom II BE ran pretty damn cool, I know my overclocked tri-core does and that supposedly pumps out well over 100W TDP!
Click to expand...

Intel allows processors to go over TDP for, I believe, up to 30 seconds at a time. But when you do this, your processor gets too hot and it will throttle itself down. That is why you can get the rollercoaster of performance and different results on the same workload.

We really don't allow TDP to be surpassed, if it is, it would only be for a couple of cycles. That is why our Turbo CORE results are more consistent.
 
JF-AMD said:
Intel allows processors to go over TDP for, I believe, up to 30 seconds at a time. But when you do this, your processor gets too hot and it will throttle itself down. That is why you can get the rollercoaster of performance and different results on the same workload.

We really don't allow TDP to be surpassed, if it is, it would only be for a couple of cycles. That is why our Turbo CORE results are more consistent.
Click to expand...

:confused:
Overclocking would go above the stated TDP, and how would a CPU get too hot and throttle itself going just above its TDP package, that'd be down to the cooler.
 
JF-AMD said:
Intel allows processors to go over TDP for, I believe, up to 30 seconds at a time. But when you do this, your processor gets too hot and it will throttle itself down. That is why you can get the rollercoaster of performance and different results on the same workload.

We really don't allow TDP to be surpassed, if it is, it would only be for a couple of cycles. That is why our Turbo CORE results are more consistent.
Click to expand...

So the higher wattage allows for higher overclocks and turbo core, rather then it being a less efficient chip?
 
JAGRocks09 said:
So the higher wattage allows for higher overclocks and turbo core, rather then it being a less efficient chip?
Click to expand...

Since when has that been the case though?.
The higher the wattage, the hotter the chip.


It'd make sense from a 95W 2.8GHZ chip to go for the 125W 3.8GHZ chip, as the extra power consumed comes at the 1GHZ clock advantage.
But this 125W chip also has a 95W variant. Like the 1055T. Guess which is the coolest and better clocking one? The 95W.
 
Martini1991 said:
Oh, I think 125W with 8 cores is "fine". But it doesn't make much sense?
Based on the cores they used in the 95W thuban, they could have released an 8 core Phenom II at 125W that would be 45nm..

I'm hearing doubt of September launch... This will be an interesting month.
Click to expand...

Not if you doubled the cache, then your 8 core phenom would suddenly want 150-170W.

Also remember, Sandybridge is on a mature and very good 32nm process, Thuban is on a very good and VERY manture 45nm process, bulldozer is on a very new and maturing process.

Theres two options really, the 125W version is either an earlier stepping, so maybe its a B2 and AMD don't want to bin millions of quid of processors, (99% of Dell buyers don't get a crap) and the 95W is a C0, or its possible you can buy ones that are essentially pre-overclocked and a 125W 3.1Ghz bulldozer will have Turbo that goes to 4.2Ghz, and a 95W 3.1Ghz chip will have turbo that goes to 3.6Ghz.
 
Martini1991 said:
It'd make sense from a 95W 2.8GHZ chip to go for the 125W 3.8GHZ chip, as the extra power consumed comes at the 1GHZ clock advantage.
But this 125W chip also has a 95W variant. Like the 1055T. Guess which is the coolest and better clocking one? The 95W.
Click to expand...
It might be OEM-only like the 95 W Phenom II X6 1055T.
 
If the quad turns out to really have a noticeably better IPC compared to the eight-core version, I might go for one of those. I'm doing so much emulating these days, IPC is everything seeing as no emulators I know of multithread very well (if at all).
 
Martini1991 said:
Wasn't B2 the fixed revision?
Click to expand...

That doesn't mean there won't be more, especially a chip on the high end on a new process. Llano might be the first 32nm chip, but its relatively low clock speed and old architecture(in parts, obviously combining them is VERY new), Bulldozer will need fixes and then likely work on steppings to improve clocks/power issues as realistically absolute priority when you get a not great chip back for testing is bug fixes over power tweaking.

Theres a lot of new stuff in Bulldozer, well, all of it, bug fixes would be first, but I'd imagine theres a lot of real world stuff to learn in terms of moving a few transitors. They've powergated each module, and its a new process with its on issues, and as GloFo improve the process a little tweak in the process could mean a slight change to the mask could end up with a decent drop in power.

I mean thats why we have 95 and 125w hexcores available, and don't forget Phenom 2's launched with 140W models, that became 125W models, that became 95W models.

For most people the power isn't an issue in the slightest and most people who buy computers have no idea how much power the whole computer uses, let alone the chip and AMD throwing away earlier steppings would be just mental.

That is the MOST likely explanation as quite simple advertising wise having two chips with the same name with different levels of Turbo is a bit, daft, complex/confusing so it would almost certainly be down to steppings.
 
It's been rumoured for ages that the first tranche will be B2 and the second tranche (at the end of this year or early January) will be B3.

That may or may not be the case.
 
You must log in or register to reply here.
Back
Top Bottom