Discussion in 'General Hardware' started by Kantastic, Oct 13, 2011.
Jeez can't even complement a guy. How you quote other forums is like apple pie to my eyes!
lol does anyone remember "real men use real cores"?
anyhow, i think bulldozer will deliver in a couple of years but amd missed the target with it for now (unless they were aiming at the server market as their primary goal)
Yes, I do. That's marketing for you, lol. How times change.
How 'bout "Real men own fabs"?
Just when I was starting to let AMD/ATI into my home.. ; ;
FX 8150 does quite well against the i7 2600k with 3 x HD 6000 series graphics cards. The key here is the i7 was 430MHz faster. The 8150 did underperform in a few games but most of the games it stood it's ground. Interesting indead....
Yeah but that introduces too many variables and limiting factors. I think in this case the memory bandwidth or PCI-E bandwidth becomes much more a limiting factor than the CPU.
Good point, never thought about that. Though it still gives you some sort of impression on performance and how it may perform for your gaming needs.
You are entirely correct.
BTW at that teaktown thingy:
Interesting, seeing how everyone is bashing DB for power consumption.
AMD took a design chance with Bulldozer, you need balls of iron to do what AMD did, to completely build something from the ground up, have your computers telling you it will perform and all of a sudden its not what was originally written on paper.
Bulldozer may very well be something of the future per say. Hopefully with enough Windows tweaks along with a few others will help boost its performance.
Though I do blame AMD for not sharing ths design with software developers years before. But then again, maybe AMD did this already, and the software developers said the hell with it. Who really knows...
Taken from here. They sure did a dandy job there! Well ok they were talking about the server chips, but still...
btw TPU needs a facepalm smiley And those 3x CFx tests look pathetic.
Ok we get it. So you promise stronger single-threaded performance when only one core of the module is loaded compared to a Thuban core and a 20% perf drop when both cores of a module are loaded. Ok that means that if an 8 core BD was made on 45nm, it would deliver more performance than a 6 core Thuban at less heat and power - yey! So we were supposed to get better performance in both instances. Yet neither of those are the case because BD is a 32nm chip, not a 45nm one! So you failed thrice! Your single threaded performance is poo, your multi-threaded performance fails to deliver and on top of that you're using what is suppposed to be a more advanced manufacturing process meaning even greater performance/power saving...Triple fail AMD!!!
Why did you release your product when you could clearly see it wasn't covering any of your promises - the ones you gave in public!
Real men don't say things with real men in them
You only sound like an ass and that is what AMD looks like now. I've been such a supporter of AMD for years but the Bulldozer fiasco is like having a sledgehammer and trying to break down the Great Wall of China. I have motherboards that can easily be upgraded to 6 core Phenom IIs and it seems like that is a much better idea than switching to a new platform. I have I7s so I won't need that anyway but it clearly seems a really bad idea to go Bulldozer now. AMD has really been in a strange place lately. They release video cards that are basically the same as the generation before in 5850, 6850 for example and they also release CPUs that are having trouble keeping up with the last generation chip. What the hell is the policy over there? Lets sacrifice today for a possible tommorrow? Good luck with that one.
Guys, guys i got it! I was so wrong! See all along I've been trying to stack BD with Thuban, but instead we should have compared it to the original Phenom!
It's coz work on BD started in 2007 and it was intended to be Phenom's replacement and and and it totally rocks compared to Phenom! See all those articles above were talking about K10 and not k10,5! And we should all stfu and instead be amazed and blown away by AMD's progress!
Well, gee I don't know...maybe because half the time BD fails to even come close to SB's performance. And please do bear in mind that AMD's mobo is supposed to deliver more PCI-E lanes and what not...And that would be ok if $ntel's solution was a 1000$ extreme edition CPU, but it's not :/
sounding like an intel fanboy thread this, bulldozer certainly isnt win at the min but its not as bad in its price point as some are making out imho, sure no ones going to be swapping from intel sandys for it but if your on a core 2 with ddr2 it dosnt look to bad if you drag mobo costs into the fray and for me at least with AMD you will have an upgrade path with intel it will be new socket time everytime apparently. just an opinion
It is interesting indeed, but of very little consequence really. Keep in mind:
- 2600k is running 500 Mhz higher while its stock clock is 200 Mhz lower. 700 Mhz more clock from stock to max OC is A LOT and surely puts the SB chip farther along the power curve, where diminishing returns kick off badly.
- Load consumption which is what most people are complaining about IS 20w higher in the BD setup.
- We would need to know exactly what both setups were doing at the time that power consumption was measured:
As stated in the power consumption pic they were benching 3dMark11, but the performance preset or the extreme one? Either way the Intel setup is producing more frames, which means higher GPU load. In the case they used the performance preset... well it would look really bad for BD. If the CPU was holding back the 3 cards like that, it was holding them back indeed, so their power consumption must of been way lower. This is why it's of critical importance the way power is measured here in TPU, not from the wall.
If this is how you feel about the test then the SB also sucks balls. Think about it, Bulldozer needs to get tweaked. Hell I am glad the thing works first of all. It was in the design face for more than 4 years. Give AMD some slack They tried something very different with Bulldozer. Intel can afford to have multiple projects running at the same time due to the massive R&D funds they have, AMD only has enough for 2 MAX.
Yes, that's true as well, if you game with multiple graphics cards, that shows you it performs competitively.
Time for a reality check, because this is obviously devolving into hate from both sides.
The thread is entitled "Why BD failed? AMD Ex-Employee speaks out!" Said title implies that the originator has a definite opinion, and thus sets the tone for the rest of the thread.
Those who love AMD, no matter what they put out, need not apply here. Poking at rabid fanboys from the other side, and then expecting anyone to listen to even reasonable statements, is like eating fast food at the zoo. It's acceptable from behind the glass, but you know that what you're doing is going to get you killed if that 800 pound gorilla decides to break the glass and steal your lunch.
Insane metaphor aside, know who you are trying to defend against. This is not the thread for people to defend AMD, this is the thread for people to bitch and moan about how AMD isn't Intel.
From the hardware agnostic, this is a thread that exists so hate doesn't spill into other threads. Please continue...
I have a 8150 right now. It's super fun to play with but aside from that it's a horrible chip. I noticed some people saying it is AMD's "Fermi". That is incorrect, it is worse. Take Fermi and decrease the performace to a HD 4870 while keeping the power usage of Fermi. That is Bulldozer. Still, like I said, it's fun to play with.
So, you know it's severe limitations, but you just had to have one to play with? That's the mark of a True PC Enthusiast.
Socket 1155 (SB) is meant to be Ivybridge compatible.
can you share a few opinion about how is working in games and multitasking? have you managed to load all cores at 100 %?
why BD faild? this is why:
BD die shot should look very much like this
but instead it looks like this
they've got 800 million transistors for in/out, logic, NB, wasted space, etc. so some how the have managed to waste a whole phenom 2 x4's worth of transistors on what should really take about, what, 150mill tansistors? also why is the "unified" L3 cache separated into 4 sections with such massive gaps?
maybe it was a learning curve using mostly software to design the chips, who knows? what is clearly evident is that they are not using the space available to them efficiently.
quoting some estimations on die area by X-bit:
"AMD publicly said that each Bulldozer dual-core CPU module with 2MB unified L2 cache contains 213 million transistors and is 30.9mm2 large. By contrast, die size of one processing engine of Llano processor (11-layer 32nm SOI, K10.5+ micro-architecture) is 9.69mm2 (without L2 cache), which indicates that AMD has succeeded in minimizing elements of its new micro-architecture so to maintain small size and production cost of the novelty.
As a result, all four CPU modules with L2 cache within Zambezi/Orochi processor consist of 852 million of transistors and take 123.6mm2 of die space. Assuming that 8MB of L3 cache (6 bits per cell) consist of 405 million of transistors, it leaves around whopping 800 million of transistors to various input/output interfaces, dual-channel DDR3 memory controller as well as various logic and routing inside the chip.
800 million of transistors - which take up a lot of die space - in an incredibly high number for various I/O, memory, logic, etc. For example, Intel's Core i-series "Sandy Bridge" quad-core chip with integrated graphics consists of 995 million."
you can't compare a diagram from a slide shot with a real core print....
spaces all are used as somehow magically all cores must communicate with each other and the other components so i don't know what's your point...
no cpu/gpu producer will leave 1 sqnm unused from a chip...
Separate names with a comma.