• 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!

Gareth

If you are the first member to review the Amd BD I suggest it would be good to have your findings in a thread of its own also, instead of being tacked to the end of this humongous rant.

just a thought :)
 
must confess, all the sites I have visited Bulldozer looks like an interesting product to try to get the best out of, there are so many clocking options. some of them give some pretty weird results as well, like the CPU-NB speed seems to do a grand total of nothing?! where-as on Phenom II if done correctly it could yield some fairly substantial gains, hell I got around ~17% increase in frames/second in Starcraft II after messing around with CPU-NB and memory timings, nice little boost in Supreme Commander as well (possibly around ~15%) but it made no difference at all in some other applications (didn't reduce performance though!). where with Bulldozer messing with it seems to do nothing, like at all? :confused:
 
gareth170 said:
i don't think so. max safe temp is 61c it was no way near that. also the clocks never droped
Click to expand...

my old Athlon 64 X2 was like that, when clocking it would give better result with a slightly lower core voltage, which at the time I thought was a bit peculiar, nothing dramatic mind. but always thought all chips had their 'sweet spot' where they perform best, voltage is just another variable to be tweaked to be honest. :)
 
Cooper said:
As posted earlier....the face behind the Derpdozer debacle...

Capturesd.png
Click to expand...

Is it me or does this guy have a bit of Peter Mandleson about him?

If so that explains a lot....
 
think people are being a bit harsh toward the guy to be fair, the 'architecture' itself is a novel and potentially powerful design, the problem is it hasn't worked out quite how they would like by fault of something inside the architecture, it isn't the idea of the Bulldozer Module that is at fault its more the implementation of that design that is the fault. so to say he should punished is a bit harsh because on paper it is a good design, and its now up to him and his team to figure out what has gone amiss with the final product, because there is one crucial part missing and that is 'holding the instructions per clock' to K10.5, if that had been the case like intended to hold fire and keep the instructions per clock the same then the dramatic increase in clock speeds combined with the multi-threaded nature of the design makes it a damned appealing prospect.

my point is don't crucify the design of the processor, just comment on how one of the fundamental points has gone wrong and see if there is a fix for the next revision, because (might be wishful thinking) if Global Foundries sort their problems out with the fabrication and AMD managed to work out the problem with instruction per clock deficit (basically bringing Bulldozer up-to its intended performance, at the moment it is costing the architecture probably in excess of 10 - 15% in performance, more in some cases) would be willing to say a lot of people here will quickly take back their pre-conceptions about the architecture. but please don't quote me on this as actual fact, there is a lot of 'maybes' in this post and I will leave it in the hands of AMD to act on the situation, just commenting on what can be done to make Bulldozer competitive as intended. put it this way if Bulldozer was my design, would be happy with the groundwork I had put in place, would go into the office, get the guys together and analyse the positives and negatives of the design, to try to understand why the instructions per clock has fallen when the original premise was to keep it at what it is or improve on it. ;)
 
I actually think that at this stage, it is too early to say if Mr Butler has done a bad job or not.

IMO, AMD need another 6-9 months to release BullDozer2. THAT CPU should be the what BD1 was supposed to be. The only problem is that by this time, Intel will probably have released their new CPU.

IMO, if AMD reduce their prices, I think they have a CPU which can sell very well (if volume sales is what AMD are after).

Now, in 6-9 months time, if BD2 is released and that also sucks...I would definitely agree with the Dunce Hat parade.

I just think that AMD (and Mr Butler) needed more time to develop the BD1. Unfortunately, due to financial reasons, AMD couldn't delay BD1 any more.
 
You must log in or register to reply here.
Back
Top Bottom