AMD Bulldozer review thread.

even games that use more than 2 cores run quite well on the i3 2100,

Bad Company 2 as one more example
bc2.gif


or GTA IV
http://www.hardware.fr/articles/778-12/grand-theft-to-iv.html


compared to a PII X2...
and looking a the PII X4 it's clear that the game scales well from 2 to 4 cores,

a combination of the amazing performance of the sandy bridge architecture in general and HT working well as can be seen on the other links I posted,


I would easily recommend a i3 2100 for someone that doesn't want to overclock and will use the PC for a more basic usage and gaming, great performance, low power usage...
 
I guess the SNB cores are so powerful that two of them + hyperthreading can equal 4 of AMD's cores.

Besides the i3-2100 I think the i5-2300 is the best gaming chip right now. It's not too expensive and fast enough at stock for everything, plus you can overclock it by 400MHz if you ever feel the need (you won't but still it's an option).

Of course in the back of my mind I'm thinking that a PII X4 955BE is only $120, the compatible mobo is a good bit cheaper, and it is still fast enough for everything and can be very easily overclocked if necessary. It's a good time to upgrade.
 
So... what do you think, would upgrading from Q6600 to Bulldozer make sense? Are those two cpus benchmarked anywhere?

A friend at work has the same dilemma and the simple and MASSIVELY unfortunate answer is - no way in hell.

I love AMD but Q6600 @ 3.2GHz (most G0 CPUs do it easily) would compete in many games with Bulldozer. It is very sad.
Even at stock speeds I would upgrade to Sct 1155 rather than AM3+ and Bulldozer.

All IMHO of course.
 
A friend at work has the same dilemma and the simple and MASSIVELY unfortunate answer is - no way in hell.

I love AMD but Q6600 @ 3.2GHz (most G0 CPUs do it easily) would compete in many games with Bulldozer. It is very sad.
Even at stock speeds I would upgrade to Sct 1155 rather than AM3+ and Bulldozer.

All IMHO of course.

A 4.5Ghz Bulldozer would equal a ~4Ghz Intel Core 2 Quad and would be a lot faster.

Only way the Q6600 would be worth keeping is if he can get it to 3.6Ghz+

My Phenom 2 x6 1075T showered scaling performance all the way up to my 24/7 daily overclock of 5Ghz.
 
Oh not you again. ;)

I wouldn't recommend Bulldozer to anyone. For the ubiquitous web browser facebook people, I'd go much cheaper, and for games SB is the only chip to consider. Assuming one wants to blow hundreds of $ on an upgrade in the first place.
 
" In fact, we're a little surprised AMD hasn't attempted to piggyback on Intel's Hyper-Threading infrastructure by making Bulldozer processors present themselves to the OS as four physical cores with eight logical threads. One would think that might be a nice BIOS menu option, at least. (Hmm. Mobo makers, are you listening?)"

Is that even possible, that the motherboard does that and kind of "bypass" the way the cpu presents itself ?
 
It sure seems like AMD should have set things up so the OS would identify every other "core" as a logical hyperthread CPU. Or perhaps the CPU was designed to ideally not need HT-style logical designations and instead we have here just more proof of problems in this implementation of the architecture.


Well I'm wondering wether it will work. You know HyperThreading share execute unit while Bulldozer don't...

What I mean is, is it easy for OS to notice the HT core?
 
Last edited by a moderator:
Well I'm wondering wether it will work. You know HyperThreading share execute unit while Bulldozer don't...

What I mean is, is it easy for OS to notice the HT core?
They do share their floating point units, and they contend for bandwidth, just like HT "virtual cores" do. I'd say Bulldozer would benefit from HT-friendly software, as long as there are less active threads than cores. If they're all loaded up, there should be no difference anymore.


---

Taking this opportunity to express my surprise at the weird cache hierarchy decisions AMD has made for BD, as well as for what appear to be signifcant empty spaces on the die.
 
good scaling, but I wonder how they tested it, I mean in terms of modules vs cores,
they set affinity manually to some cores, or limit the game somehow?
if they set affinity, they gave priority to the same modules, or different modules (for 4 and less)?

anyway, BF3 doesn't seem hard on CPUs particularly on these single player tests, they should try to test it on the 64players mp maps, but of course is a bit harder to test accurately,
 
Yeah 64p multiplayer is way more CPU intensive. But like you said it would be practically impossible to test accurately.

Maybe all the review sites should get together and create a Benchmark Server where they all run around and to the exact same things over and over :p
 
does the drop. of 3dnow could've cause the low FP performance (compared to the phenom)

no.

1) 3dnow did not support all required rounding modes, it was not if ieee-754 compatible. It was not used a lot for "general FP calculations", compilers could not use it without some extra "I know what I am doing, allow using them even if it breaks standards"- flags
2) 3dnow only supports 32-bit single-precision-numbers, not 64-bit double-precision numbers.
3) Everything that 3dnow could do, SSE did better and faster. Already K10 could do SSE twice faster than 3dnow, and if bulldozer would have supported 3dnow, the 3dnow performance would have been half of SSE performance.
4) In cases where bulldozer has slower FPU performance than phenom, it's because Bulldozer has less FPU units; one module has only one FPU, shared with two bulldozer "cores", but every phenom core has it's own FPU.
 
Yeah 64p multiplayer is way more CPU intensive. But like you said it would be practically impossible to test accurately.

Maybe all the review sites should get together and create a Benchmark Server where they all run around and to the exact same things over and over :p

I thought the article on Tom's that implied even a single, hyperthreaded core was adequate for BF3 incredibly irresponsible. All based on a 90 second single player sequence? I think testing MP would be doable, the problem is the time commitment. You'd want to play a number of hours on an active 64 player server with each CPU to get enough data for an average that'd be comfortably comparable.
 
A 4.5Ghz Bulldozer would equal a ~4Ghz Intel Core 2 Quad and would be a lot faster.

To be frank, this is rather bullshit(y), and I am somewhat opposed to bullshit. Therefore, for future reference, I will politely urge you to invest more thought in your posts, as this sort of wild partisanship doesn't help in general, nor does it have a place on B3D. Thank you.
 
At a 4.5ghz base clock, BD would probably look good compared to Phenom II and C2Q, but overclockers are finding it to be stupidly hot. Like it makes Prescott look good. 4.5 ghz is the clock speed I think the thing needed to ship at but obviously it wasn't gonna happen with its power usage.

It would be rather pointless to upgrade from a 4ghz Yorkfield to it. Throwing money away.
 
To be frank, this is rather bullshit(y), and I am somewhat opposed to bullshit. Therefore, for future reference, I will politely urge you to invest more thought in your posts, as this sort of wild partisanship doesn't help in general, nor does it have a place on B3D. Thank you.

:rolleyes:

Core 2 Quad and Phenom 2 are pretty neck and neck in terms of per-clock performance and BD is 10-15% slower per-clock then Phenom 2.... So get a clue tbh as it's not BS..

And my post was relevent to the point in hand..... You don't like it, Don't read it...simple.. :rolleyes:
 
Back
Top