InternationalHat
[H]ard|Gawd
- Joined
- Aug 13, 2004
- Messages
- 1,481
I think they should have toned down the codename to give us a better understanding of the performance like instead of AMD Bulldozer maybe AMD Donkeycart.
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
AMD Responds
- Thread starter Ryan45
- Start date
jojo69
[H]F Junkie
- Joined
- Sep 13, 2009
- Messages
- 11,267
old saying, be greedy when others are fearful, fearful when others are greedy
perhaps this is the time to snatch up a bargain AM3+ MOBO and wait for a few steppings and see what happens
perhaps this is the time to snatch up a bargain AM3+ MOBO and wait for a few steppings and see what happens
needmorecarnitine
Supreme [H]ardness
- Joined
- Feb 10, 2004
- Messages
- 4,609
the Wheelbarrow
needmorecarnitine
Supreme [H]ardness
- Joined
- Feb 10, 2004
- Messages
- 4,609
Wat?
That would applies to something like stocks/equities
Why would you buy a "bargain" AM3+ motherboard now when its going to take months for AMD to improve this situation (if it ever does). Do you actually think AM3+ motherboards are going to go up in price?
jojo69
[H]F Junkie
- Joined
- Sep 13, 2009
- Messages
- 11,267
well yeah, from reading on here seems like you can't give them away
__hollywood|meow
[H]ard|Gawd
- Joined
- Feb 20, 2006
- Messages
- 1,500
are you kidding? wat could he even say that wouldnt be met by endless flaming?
ordovician
2[H]4U
- Joined
- Dec 28, 2007
- Messages
- 2,625
Here's to hoping a magic stepping revision will give them 5ghz+. lol. Thoroughred-B anyone? 
hey, we should all take it easy on the AMD marketing team.
they didn't have much to work with. It's hard to market a turd. You try it.
Anyway, for all the core-whores, Intel should really make a 24-core Atom.
So many coars.. all your VM's are belong to us
they didn't have much to work with. It's hard to market a turd. You try it.
Anyway, for all the core-whores, Intel should really make a 24-core Atom.
So many coars.. all your VM's are belong to us
yes i'm pretty sure AMD's marketing team want to kill themselves this week
poor bastards...
bboynitrous
2[H]4U
- Joined
- Feb 29, 2004
- Messages
- 2,514
I was thinking the same. I hoping Bulldozer develops into something more down the road. I mean it's kind of sad to upgrade to a new processor and go AMD still when my Phenom II x6 is maybe slightly slower than Bulldozer. I'm really not going to upgrade to that. I wouldn't, as a new system builder, consider going Bulldozer when you could either go budget Phenom II now, or Sandy Bridge which is the best bang for your buck.
Who knows. This one took 3 years.
You know its funny, reading the posts of some folks on here. Well no its not funny. They write as though some great personal tragedy has occurred to them. Some great injustice or harm has befallen humanity.
Has it? No not really. A company most folks have never heard of released an average CPU that will do much the same as most CPUs, maybe not as fast as some but faster than others. Oh and no one really needs it as a matter of life or death either.
That's it really.
Talk about some folks seriously need a b**w job!
Has it? No not really. A company most folks have never heard of released an average CPU that will do much the same as most CPUs, maybe not as fast as some but faster than others. Oh and no one really needs it as a matter of life or death either.
That's it really.
Talk about some folks seriously need a b**w job!
Geez that's a ways away. Prob would have better luck crossing fingers for a MC price drop huh?
I just built a system with 2500k. I got the processor for cheap.
It amazes me that people are crying about this, is anyone going to see a difference on any given day? Are you going to be able to pin point it. If you are encoding, are you encoding with single thread. Yes it is not that fast, yes the processor disappointed many, but in reality would you be running this processor and encoding with apps that are single threaded, you are getting eight cores, that will probably be up to par with the 2600k when used in heavily threaded apps or may be even faster.
People will buy bulldozer regardless of the select online gamer community, the processor doesn't seem to bad when you are gaming, AMD wants to get you high on 8 cores, thats what they are selling you, if you dont need them then buy the quad core and quit crying.
It amazes me that people are crying about this, is anyone going to see a difference on any given day? Are you going to be able to pin point it. If you are encoding, are you encoding with single thread. Yes it is not that fast, yes the processor disappointed many, but in reality would you be running this processor and encoding with apps that are single threaded, you are getting eight cores, that will probably be up to par with the 2600k when used in heavily threaded apps or may be even faster.
People will buy bulldozer regardless of the select online gamer community, the processor doesn't seem to bad when you are gaming, AMD wants to get you high on 8 cores, thats what they are selling you, if you dont need them then buy the quad core and quit crying.
Well from what i have read and seen of the technical specs for BD, its faults lie in its cache system. First mistake was having the modules on the cores share from a single cache. second mistake was actually reducing the size of the cache from 64kb to 16kb. many of the reviews show that the module cache sharing has a very large performance impact of about 40%. the testers got a higher performance scaling when they used 1 module from each core rather than using both modules from the first and second core core. this was done on a 2 core test. A visual example of this would be:
C=core and M=module
M1+M2 of C1 was much slower that using M1 of C1 + M1 of C2. so yeah this shows that the fault for BD performance is the reduced cache size along with the shared cache system.
In short the cache system is choking the data through put of the CPU cores.
The reason why it feels like a phenom 2 x6 is because it has the through put of one if even less than that.
This is a snippet from anandtechs review:
Cache Hierarchy and Memory Subsystem
Each integer core features its own dedicated L1 data cache. The shared FP core sends loads/stores through either of the integer cores, similar to how it works in Phenom II although there are two integer cores to deal with now instead of just one. Bulldozer enables fully out-of-order loads and stores, an improvement over Phenom II putting it on parity with current Intel architectures. The L1 instruction cache is shared by the entire bulldozer module, as is the L2 cache.
The instruction cache is a large 64KB 2-way set associative cache, similar in size to the Phenom II's L1 cache but obviously shared by more "cores". A four-core Phenom II would have 256KB of total L1 I-Cache, while a four core Bulldozer will have half of that. The L1 data caches are also significantly smaller than Bulldozer's predecessor. While Phenom II offered a 64KB L1 D-Cache per core, Bulldozer only offers 16KB per integer core.
But here is the whole review for those who want too read it: http://www.anandtech.com/show/4955/the-bulldozer-review-amd-fx8150-tested/6
C=core and M=module
M1+M2 of C1 was much slower that using M1 of C1 + M1 of C2. so yeah this shows that the fault for BD performance is the reduced cache size along with the shared cache system.
In short the cache system is choking the data through put of the CPU cores.
The reason why it feels like a phenom 2 x6 is because it has the through put of one if even less than that.
This is a snippet from anandtechs review:
Cache Hierarchy and Memory Subsystem
Each integer core features its own dedicated L1 data cache. The shared FP core sends loads/stores through either of the integer cores, similar to how it works in Phenom II although there are two integer cores to deal with now instead of just one. Bulldozer enables fully out-of-order loads and stores, an improvement over Phenom II putting it on parity with current Intel architectures. The L1 instruction cache is shared by the entire bulldozer module, as is the L2 cache.
The instruction cache is a large 64KB 2-way set associative cache, similar in size to the Phenom II's L1 cache but obviously shared by more "cores". A four-core Phenom II would have 256KB of total L1 I-Cache, while a four core Bulldozer will have half of that. The L1 data caches are also significantly smaller than Bulldozer's predecessor. While Phenom II offered a 64KB L1 D-Cache per core, Bulldozer only offers 16KB per integer core.
But here is the whole review for those who want too read it: http://www.anandtech.com/show/4955/the-bulldozer-review-amd-fx8150-tested/6
Last edited:
Your guys' knowledge of cores, cache, pipes, and instructions is far beyond mine. All I know is that my 2 AMD computers (both AM3, flashable to AM3+) will NOT be going bulldozer. My 1090T is better, for my purposes, than anything I've seen about BD on the reviews. My 720x3 is hurting. I'll probably go 1100T to replace that one. (My 2 intel machines will continue to stay intel. I split my builds: I want competition for my money.)
C'mon AMD.
I'll probably go 1100T to replace that one. (My 2 intel machines will continue to stay intel. I split my builds: I want competition for my money.)C'mon AMD.
Correction. The Bulldozer project took 7 years. So the turd is well baked.
Indeed. If they want to see which one is the baddest gaming processor one has to take GPU out of the equation, meaning dropping resolution and details to lowest possible and then measure FPS.
Which leads me to why HardOCP hasnt tested Bulldozer this way yet?
NIVO
[H]ard|Gawd
- Joined
- Jul 19, 2004
- Messages
- 1,253
ok, AMD, based off your account of BF3 you won by 1-3 fps! Yay? Nay nay, I say "FAIL"
buttons
2[H]4U
- Joined
- Oct 12, 2011
- Messages
- 2,174
bomberboysk
Gawd
- Joined
- Jun 7, 2009
- Messages
- 760
Netburst rocked?
It did?
Riccochet
Fully [H]
- Joined
- Apr 11, 2007
- Messages
- 29,906
Astheruinsfall
n00b
- Joined
- Oct 4, 2010
- Messages
- 5
As a space heater maybe.. heat and power consumption were off the charts.
Not to mention it only got to even decent performance after multiple process shrinks and significant redesigns, something AMD absolutely cannot do.
JayteeBates
[H]ard|Poof
- Joined
- Jul 21, 2007
- Messages
- 5,104
When you have to "explain" your newly released product to the intended market you are already behind.
While I'm not saying this part isn't true (who couldn't use more of those?), the fact of the matter is that people who hang out at a place like [H] are typically enthusiasts who are passionate about the craft. I don't think even Intel fanboys wanted to see AMD fail this badly, so of course people here are going to have something to say about it. The thing that's sad about it is that despite years on the design board, their best design doesn't even beat their own previous offering released more than 1.5 years ago on a consistent basis. The issue is not only being non-competetive, it's also being stagnant on their own performance.
Server chip? Heat and power usage are very important factors in a data center, where an inefficient CPU like BD doesn't belong. I wouldn't even replace a C2D-based Xeon server with a BD server. It is the absolute last thing on my list. Refurb Nehalem-based servers are a dime a dozen and will surely be far cheaper than new BD servers.
After 10 years on [H], I really hope you are joking.... P4 rocked so hard that it spent a lot of its life being outperformed by Athlon64, and that the Intel design team went back to the Pentium 3 architecture to start developing their Core and Core2 designs.....
But... you do have a point in the fact that P4 was worse than P3 starting out but ultimately outperformed it later.
Intel never stopped developing that architecture. The Pentium M was in between the P3 and the first Core. I remember having an earlier P4 and a Pentium M at the same time and wondering why the Pentium M actually felt faster. It was simple: it actually was faster.
burningrave101
[H]F Junkie
- Joined
- Sep 9, 2003
- Messages
- 11,825
The P4 and Athlon64 traded blows for the most part with the Athlon64 excelling more in areas where latency is important such as gaming due to it having an on-die memory controller. The higher clock speeds and HyperThreading however kept the P4 ahead in multimedia with encoding and rendering and more work related tasks along with multitasking performance.
Bulldozer doesn't have that advantage in clock speeds over Sandy Bridge though and AMD's 8 core design so far isn't showing to be any better than Intel's HyperThreading.
Probably because no one actually plays like that. Who cares if it is faster at 640x480?
And neither does anyone run benchmarks all day long, jerking to numbers they produce. Or dear God I atleast hope not.
Point is that this method is just a form of benchmarking, measuring CPUs "raw horsepower" in gaming enviroment. CPU's potential in case of heavier games in future?
Maybe so we know which ones is the more powerful chip?
If two chips cost the same, but one does 60 FPS, and the other does 90 FPS, which one do you think will last longer as time progresses and programs/games become more demanding?
More powerful chip at pointless use cases isn't much of an indicator. Just because it can run Quake at 1500FPS or Metro 2033 at 640x480 at 100FPS doesn't say anything about how it's going to run tomorrow's games - because no one knows how tomorrow's games are going to scale - and it doesn't tell us anything about how it runs today's games because no one runs games at those resolutions. It's the same philosophy of why [H] doesn't test video cards at low resolutions - because no one play them that way, so the information you are providing via those benchmarks isn't useful.
If you have unlimited time and resources then of course, test every possible iteration - but in a resource constrained (real-world) environment, test the use cases that people actually, you know, use.