The CPU does very well in multi threaded situations, and should serve its purpose in tackling the server space.
The Bulldozer core was never specifically designed to go after SandyBridge in gaming or single threaded apps.
| Forum Index > General Forum |
|
IreScath
Canada521 Posts
The CPU does very well in multi threaded situations, and should serve its purpose in tackling the server space. The Bulldozer core was never specifically designed to go after SandyBridge in gaming or single threaded apps. | ||
|
da_head
Canada3350 Posts
![]() really disappointed in amd though, a monopolized market is never good :S | ||
|
JingleHell
United States11308 Posts
On October 13 2011 01:34 B00ts wrote: The Bulldozer core is and was designed for the Server space. All they did was alter some features and throw an FX brand on it and labelled it their Consumer CPU 'top of the line'. The CPU does very well in multi threaded situations, and should serve its purpose in tackling the server space. The Bulldozer core was never specifically designed to go after SandyBridge in gaming or single threaded apps. No, it does a functional job in multi-threaded situations. Or is "Slightly better than half the physical cores in an older CPU sometimes" equivalent to "does very well" now? | ||
|
Rannasha
Netherlands2398 Posts
Now with the OEMs being bullied away from AMD by Intel, AMD never got the market share and revenues to expand their R&D division. I fear for competition on the CPU market if AMD can't come back with a strong processor soon. At least they're making a profit now, something that wasn't true a few years back. But still... | ||
|
IreScath
Canada521 Posts
On October 13 2011 01:38 JingleHell wrote: Show nested quote + On October 13 2011 01:34 B00ts wrote: The Bulldozer core is and was designed for the Server space. All they did was alter some features and throw an FX brand on it and labelled it their Consumer CPU 'top of the line'. The CPU does very well in multi threaded situations, and should serve its purpose in tackling the server space. The Bulldozer core was never specifically designed to go after SandyBridge in gaming or single threaded apps. No, it does a functional job in multi-threaded situations. Or is "Slightly better than half the physical cores in an older CPU sometimes" equivalent to "does very well" now? Yes, look at the prices of an i7 vs and the FX as well. I would rather by a CPU that does better (albeit marginally small) for less money for my server space. That, and Win7 is not optomized when it comes to thread placement either, and we've only seen limited Win8 tests, however thread scheduling is addressed (or so I hear). Also, I said 'does very well'. If an i7 does very well, and teh FX is just as good (above or below a few % in different tests), that means it also does very well.[DISCLAIMER]: I am talking about multi-threaded tests ONLY.[/DISCLAIMER] Don't nit-pick my choice of words and grammar... You know darn well what I was saying. And that when people say "does very well", they are not saying "does totally better than competition"... Otherwise they would just say that, if that is what they meant. | ||
|
JingleHell
United States11308 Posts
On October 13 2011 01:46 B00ts wrote: Show nested quote + On October 13 2011 01:38 JingleHell wrote: On October 13 2011 01:34 B00ts wrote: The Bulldozer core is and was designed for the Server space. All they did was alter some features and throw an FX brand on it and labelled it their Consumer CPU 'top of the line'. The CPU does very well in multi threaded situations, and should serve its purpose in tackling the server space. The Bulldozer core was never specifically designed to go after SandyBridge in gaming or single threaded apps. No, it does a functional job in multi-threaded situations. Or is "Slightly better than half the physical cores in an older CPU sometimes" equivalent to "does very well" now? Yes, look at the prices of an i7 vs and the FX as well. I would rather by a CPU that does better (albeit marginally small) for less money for my server space. That, and Win7 is not optomized when it comes to thread placement either, and we've only seen limited Win8 tests, however thread scheduling is addressed (or so I hear). Also, I said 'does very well'. If an i7 does very well, and teh FX is just as good (above or below a few % in different tests), that means it also does very well. Don't nit-pick my choice of words and grammar... You know darn well what I was saying. And that when people say "does very well", they are not saying "does totally better than competition"... Otherwise they would just say that, if that is what they meant. Who's nitpicking? I'm trying to understand how a newer CPU with a higher clock and twice the physical cores occasionally barely exceeding the competition is considered doing very well. It's counter intuitive. You can try to make it sound good all you want, but if a CPU needs twice the cores and a higher clock to match performance on apps that use the cores even reasonably well, it just plain sucks. Play fanboy all you want, but don't expect us to buy into this line of... uhm... thinking. | ||
|
IreScath
Canada521 Posts
On October 13 2011 01:50 JingleHell wrote: Show nested quote + On October 13 2011 01:46 B00ts wrote: On October 13 2011 01:38 JingleHell wrote: On October 13 2011 01:34 B00ts wrote: The Bulldozer core is and was designed for the Server space. All they did was alter some features and throw an FX brand on it and labelled it their Consumer CPU 'top of the line'. The CPU does very well in multi threaded situations, and should serve its purpose in tackling the server space. The Bulldozer core was never specifically designed to go after SandyBridge in gaming or single threaded apps. No, it does a functional job in multi-threaded situations. Or is "Slightly better than half the physical cores in an older CPU sometimes" equivalent to "does very well" now? Yes, look at the prices of an i7 vs and the FX as well. I would rather by a CPU that does better (albeit marginally small) for less money for my server space. That, and Win7 is not optomized when it comes to thread placement either, and we've only seen limited Win8 tests, however thread scheduling is addressed (or so I hear). Also, I said 'does very well'. If an i7 does very well, and teh FX is just as good (above or below a few % in different tests), that means it also does very well. Don't nit-pick my choice of words and grammar... You know darn well what I was saying. And that when people say "does very well", they are not saying "does totally better than competition"... Otherwise they would just say that, if that is what they meant. Who's nitpicking? I'm trying to understand how a newer CPU with a higher clock and twice the physical cores occasionally barely exceeding the competition is considered doing very well. It's counter intuitive. You can try to make it sound good all you want, but if a CPU needs twice the cores and a higher clock to match performance on apps that use the cores even reasonably well, it just plain sucks. Play fanboy all you want, but don't expect us to buy into this line of... uhm... thinking. Who is a fanboy? And why all the attacks? I was merely pointing out the fact that people are getting a little too upset over the crappy results in some tests... as the chip wasn't really made for it... and also isn't a pure Consumer-designed CPU. Also... It doesn't really have twice the physical cores. The 8 core has 4 'modules'. Each module has a floating point and 2 integer cores... So its sort of in between 4 and 8 cores, as a regular core has 1 of each. Im not making excuses for them or anything... just clarifying. And your definition of "uses cores reasonably well" is of course, subjective. | ||
|
Antisocialmunky
United States5912 Posts
| ||
|
Crying
Bulgaria778 Posts
WTF 600WATT Processor in LOAD With clock???My whole system draws 550W and thats my whole PSU,it never draws more than 400-450W,even thou i have 4ghz overclock on my 32nm i3. This is a complete joke. | ||
|
Antisocialmunky
United States5912 Posts
On October 13 2011 01:58 Crying wrote: Intel is so far ahead in the technological aspect related to CPU's. WTF 600WATT Processor in LOAD With clock???My whole system draws 550W and thats my whole PSU,it never draws more than 400-450W,even thou i have 4ghz overclock on my 32nm i3. This is a complete joke. Its 600Watt system draw. BD has a similar draw to the old i7's. | ||
|
Rannasha
Netherlands2398 Posts
On October 13 2011 01:50 JingleHell wrote: Show nested quote + On October 13 2011 01:46 B00ts wrote: On October 13 2011 01:38 JingleHell wrote: On October 13 2011 01:34 B00ts wrote: The Bulldozer core is and was designed for the Server space. All they did was alter some features and throw an FX brand on it and labelled it their Consumer CPU 'top of the line'. The CPU does very well in multi threaded situations, and should serve its purpose in tackling the server space. The Bulldozer core was never specifically designed to go after SandyBridge in gaming or single threaded apps. No, it does a functional job in multi-threaded situations. Or is "Slightly better than half the physical cores in an older CPU sometimes" equivalent to "does very well" now? Yes, look at the prices of an i7 vs and the FX as well. I would rather by a CPU that does better (albeit marginally small) for less money for my server space. That, and Win7 is not optomized when it comes to thread placement either, and we've only seen limited Win8 tests, however thread scheduling is addressed (or so I hear). Also, I said 'does very well'. If an i7 does very well, and teh FX is just as good (above or below a few % in different tests), that means it also does very well. Don't nit-pick my choice of words and grammar... You know darn well what I was saying. And that when people say "does very well", they are not saying "does totally better than competition"... Otherwise they would just say that, if that is what they meant. Who's nitpicking? I'm trying to understand how a newer CPU with a higher clock and twice the physical cores occasionally barely exceeding the competition is considered doing very well. It's counter intuitive. You can try to make it sound good all you want, but if a CPU needs twice the cores and a higher clock to match performance on apps that use the cores even reasonably well, it just plain sucks. Play fanboy all you want, but don't expect us to buy into this line of... uhm... thinking. I don't think you can simply compare the number of cores and clockspeed and come up with a performance expectation. The variables that matter are price and possibly power consumption. There are various roads that lead to good performance. That more clockspeed doesn't necessarily mean better performance we already knew from the Pentium 4. Similarly, more cores doesn't automatically mean better performance (even on multi-threaded programs) as the individual cores can be rather weak, as we see now with Bulldozer. Compare performance for the set of applications/tasks that you're interested in and match that with the price. Pulling in other statistics such as clockspeed, core count, amount of L3 cache, whatever, is fairly pointless. Also note that despite AMDs marketing buzz, Bulldozer isn't a full 8-core CPU. It consists of 4 so-called "modules" that each contain almost 2 cores. Almost meaning that some of the parts, most notably the floating point unit and some of the cache, are shared between the 2 cores in one module. This implementation lies between Intels HyperThreading and having 8 actually distinct cores both in terms of performance and amount of transistors required. | ||
|
Bibdy
United States3481 Posts
| ||
|
Antisocialmunky
United States5912 Posts
On October 13 2011 02:01 Bibdy wrote: Well, where's the test that shows off its merits? If it was intended for a specific niche, shouldn't that be the benchmark? I find it more likely that they hedged a lot of bets on a single research line that was not particularly fruitful, and they've decided to release something to try and get some of that investment back. Well, you'll get those with the server parts since only very few desktop things can use BD's power. | ||
|
MangoTango
United States3670 Posts
![]() Intel has been wiping the floor with AMD for a while now. | ||
|
IreScath
Canada521 Posts
On October 13 2011 02:01 Bibdy wrote: Well, where's the test that shows off its merits? If it was intended for a specific niche, shouldn't that be the benchmark? I find it more likely that they hedged a lot of bets on a single research line that was not particularly fruitful, and they've decided to release something to try and get some of that investment back. If you follow the industry at all... You would know that the new platform was designed for The Server market. However, I'm fully aware that not everyone is as nerdy as I when it comes to this stuff... But any google search for Bulldozer will eventually get you search results from prior to today and you can plainly see this fact. ![]() | ||
|
JingleHell
United States11308 Posts
On October 13 2011 02:01 Bibdy wrote: Well, where's the test that shows off its merits? If it was intended for a specific niche, shouldn't that be the benchmark? I find it more likely that they hedged a lot of bets on a single research line that was not particularly fruitful, and they've decided to release something to try and get some of that investment back. Sure, except they've tried to market it for all the things it's being shredded for not doing well. Market an 8 core CPU, I treat bench results like an 8 core CPU. Tough shit for AMD on that, one of their marketing points just makes the performance look worse. And I know full well that clock and cores aren't the primary components to performance. The problem is when you release a CPU that does a miserable job at a lot of the things you tried to market it for, even with edges in things like that. And anybody acting like BD is worth a damn outside incredibly limited areas, just like Ph2, is deluding themselves. Obviously nobody will say it's 100% useless for everything, but they've fallen so far short of so many of the marks they set for themselves and hyped, that it's a huge disappointment. | ||
|
Bibdy
United States3481 Posts
On October 13 2011 02:05 B00ts wrote: Show nested quote + On October 13 2011 02:01 Bibdy wrote: Well, where's the test that shows off its merits? If it was intended for a specific niche, shouldn't that be the benchmark? I find it more likely that they hedged a lot of bets on a single research line that was not particularly fruitful, and they've decided to release something to try and get some of that investment back. If you follow the industry at all... You would know that the new platform was designed for The Server market. However, I'm fully aware that not everyone is as nerdy as I when it comes to this stuff... But any google search for Bulldozer will eventually get you search results from prior to today and you can plainly see this fact. ![]() I'm learning about processors, memory, memory management, caches yada yada in one my classes right now, so I'm genuinely interested in the stuff. So, literally just googled "AMD Bulldozer target market" and it's not 'clear' whatsoever. Sounds like one needs to do some serious digging to find that kind of relevant information. I mean, if it was so blatantly obvious, don't you think the guys testing the thing would throw it into a webserver and seeing how it performs under load? Did that little nugget of information just completely pass over everyone's head? Or is that just an excuse? After more website-skimming, I'm seeing a lot of quotes along the lines of What's interesting to note is that the Bulldozer architecture is being launched for both the server and desktop markets. If that's true, then the desktop people are disappointed, and rightfully so, I believe. So, how does the thing perform as a server? Does it blow the competition out of the water? | ||
|
IreScath
Canada521 Posts
On October 13 2011 02:09 Bibdy wrote: Show nested quote + On October 13 2011 02:05 B00ts wrote: On October 13 2011 02:01 Bibdy wrote: Well, where's the test that shows off its merits? If it was intended for a specific niche, shouldn't that be the benchmark? I find it more likely that they hedged a lot of bets on a single research line that was not particularly fruitful, and they've decided to release something to try and get some of that investment back. If you follow the industry at all... You would know that the new platform was designed for The Server market. However, I'm fully aware that not everyone is as nerdy as I when it comes to this stuff... But any google search for Bulldozer will eventually get you search results from prior to today and you can plainly see this fact. ![]() If that's true, then the desktop people are disappointed, and rightfully so, I believe. So, how does the thing perform as a server? Does it blow the competition out of the water? I haven't seen / can't find anything yet. Anyone else? | ||
|
Cocoabean
Canada90 Posts
On October 13 2011 01:57 B00ts wrote: I was merely pointing out the fact that people are getting a little too upset over the crappy results in some tests... as the chip wasn't really made for it... and also isn't a pure Consumer-designed CPU. So AMD spent the past 5 years and all their R+D designing a chip that is tailored only for a super-specific niche computer market? Lets not kid ourselves here. They released an absolute turd of a product. When the i7-2600k, which has been out for almost close to a year now, completely annihilates the FX-8150 in virtually every benchmark that is applicable to 99.9% of the consumer base out there, you know AMD has a massive failure on their hands. To justify this chip is like saying Honda released a new Accord at the same price as a new Camry, but has inferior performance in every single driving scenario possible except if you drive exactly at 87mph on gravel road in downtown San Francisco on October 27th of 2013. Again, AMD managed to do the impossible and actually released *significantly* inferior CPU performance to those that have already been out on the market for 8 months-2 years. The fact that this lineup of CPU's does decently better at a completely niche market is irrelevant. | ||
|
IreScath
Canada521 Posts
On October 13 2011 02:15 Cocoabean wrote: Show nested quote + On October 13 2011 01:57 B00ts wrote: I was merely pointing out the fact that people are getting a little too upset over the crappy results in some tests... as the chip wasn't really made for it... and also isn't a pure Consumer-designed CPU. So AMD spent the past 5 years and all their R+D designing a chip that is tailored only for a super-specific niche computer market? Servers are super niche? lol Lets not kid ourselves here. They released an absolute turd of a product. When the i7-2600k, which has been out for almost close to a year now, completely annihilates the FX-8150 in virtually every benchmark that is applicable to 99.9% of the consumer base out there, you know AMD has a massive failure on their hands. This is fair... Except that the FX-8150 is not ~$314, but $245. Still I think the price/performance ratio leans towards Intel still. To justify this chip is like saying Honda released a new Accord at the same price as a new Camry, but has inferior performance in every single driving scenario possible except if you drive exactly at 87mph on gravel road in downtown San Francisco on October 27th of 2013. Again, AMD managed to do the impossible and actually released *significantly* inferior CPU performance to those that have already been out on the market for 8 months-2 years. The fact that this lineup of CPU's does decently better at a completely niche market is irrelevant. I had no idea servers were niche... hmm Actually, when it comes to CPU sales... Do you know what the market is actually like right now? I'll give you a hint... Gaming PC's are niche. | ||
| ||
StarCraft 2 StarCraft: Brood War Sea Dota 2Hyuk Free BeSt Killer Zeus EffOrt Sharp ToSsGirL Aegong [ Show more ] League of Legends Counter-Strike Other Games Organizations Dota 2 Other Games StarCraft: Brood War StarCraft 2 StarCraft: Brood War
StarCraft 2 • Berry_CruncH102 StarCraft: Brood War• LUISG • AfreecaTV YouTube • intothetv • Kozan • IndyKCrew • LaughNgamezSOOP • Migwel • sooper7s League of Legends |
|
The PondCast
Replay Cast
RSL Revival
herO vs Zoun
Classic vs Reynor
Maru vs SHIN
MaxPax vs TriGGeR
BSL: GosuLeague
RSL Revival
WardiTV Korean Royale
RSL Revival
WardiTV Korean Royale
IPSL
Julia vs Artosis
JDConan vs DragOn
RSL Revival
[ Show More ] Wardi Open
IPSL
StRyKeR vs OldBoy
Sziky vs Tarson
Replay Cast
Monday Night Weeklies
Replay Cast
Wardi Open
|
|
|