Overclock.net banner

1 - 20 of 136 Posts

·
Registered
Joined
·
636 Posts
I'll believe it when I OC it
biggrin.gif
 

·
Premium Member
Joined
·
5,180 Posts
Quote:
Originally Posted by Seronx View Post

Or as other people have said MOAR THREADZ
http://www.legitreviews.com/article/1838/1/
In other news
The AMD Reality Check Results:
System C (Intel Core i7-2700K): 40 Votes
System D (AMD FX-8150): 73 Votes
No Difference: 28 Votes
FX-8150 won
thumb.gif
With AMD sponsoring the test.
 

·
Frequency is Megabytes
Joined
·
2,997 Posts
Quote:
Originally Posted by trumpet-205 View Post

With AMD sponsoring the test.
It was a blind test....
rolleyes.gif


But, I guess some people can't read
Quote:
Originally Posted by AMD4ME View Post

This information has been known for awhile. It was taken from the AMD slide presentation to their OEMs.
http://www.xbitlabs.com/news/cpu/display/20120116163742_AMD_Describes_Piledriver_Architecture_Peculiarities_to_Software_Developers.html
I know I even talked about it in my Piledriver thread in AMD CPUs and my FD-SOI thread in Rumours
 

·
Premium Member
Joined
·
4,740 Posts
I've seen that chart around before. Also the rumor community has been talking about 10 cores since before Zambezi came out.

The just as if not more important news here though is the 64kB L1 cache per module, new instructions, and quad-channel memory. Considerably higher IPC and AMD finally having a chance at catching up to Intel (mainstream chipsets at least) in memory bandwidth? Probably more important than 25% more cores that most people probably won't benefit much from until software finally catches up.
Quote:
Originally Posted by Seronx View Post

Or as other people have said MOAR THREADZ
http://www.legitreviews.com/article/1838/1/
In other news
The AMD Reality Check Results:
System C (Intel Core i7-2700K): 40 Votes
System D (AMD FX-8150): 73 Votes
No Difference: 28 Votes
FX-8150 won
thumb.gif
I'm surprised. Either the test and/or the results are illegitimate or... umm... nice job. But we all know that if there were two or three 7970s, the results would probably be much different.
 

·
Frequency is Megabytes
Joined
·
2,997 Posts
Quote:
Originally Posted by trumpet-205 View Post

Yes, it was a blind test, but systems were prepared by AMD. How exactly is it a fair test?
They had the same specs except the motherboard and CPU how exactly is that rigged?
Quote:
Originally Posted by TFL Replica View Post

Double the cores every 2 years. They have no choice.
applaud.gif
They aren't doubling the cores...every two years

Bulldozer is 8 cores,
Piledriver is 10 cores,(Early Piledriver is 4 cores)
Steamroller is 10 cores
Excavator will most likely be 12 cores but might still be 10 cores(2009 -> 2014 isn't 2 years)
Quote:
Originally Posted by Usario View Post

I'm surprised. Either the test and/or the results are illegitimate or... umm... nice job. But we all know that if there were two or three 7970s, the results would probably be much different.
So, you are saying Bulldozer is faster with single GPUs?
handlebarssmiley.png
 

·
Premium Member
Joined
·
5,180 Posts
Quote:
Originally Posted by Seronx View Post

They had the same specs except the motherboard and CPU how exactly is that rigged?
What about software wise? This test will be more believable if it is sponsored by a third-party group where Intel and AMD are not involved.

From statistical standpoint this is a biased test (Sorry).
 

·
Frequency is Megabytes
Joined
·
2,997 Posts
Quote:
Originally Posted by trumpet-205 View Post

What about software wise? This test will be more believable if it is sponsored by a third-party group where Intel and AMD are not involved.
They were video games...rofl
Quote:
Originally Posted by trumpet-205 View Post

From statistical standpoint this is a biased test (Sorry).
Really now...even when you add the unchanged to the i7 2700K, the FX8150 still wins

System C (Intel Core i7-2700K): 40 Votes, 40+28 = 68
System D (AMD FX-8150): 73 Votes
No Difference: 28 Votes
 

·
Enthusiast Gamer
Joined
·
3,829 Posts
Will Piledriver be faster than BD. Absolutely. However will it be faster or equal to Sandy/Ivy. I highly doubt it. AMD has struggled behind Intel ever since Core2Due arrived and It will take a miracle for AMD to become #1 again. Until they nail Core efficiency they will never beat Intel. I don't care if they release a 10 Core Processor it still won't be faster because Intel will just get even more efficient.

I mean look at the i3 2100 vs a Phenom II 965 BE. The i3 slightly edges out the 965BE, but one is a dual core(i3) and one is a quad core(965BE). Just imagine How good Haswell will be.
 
  • Rep+
Reactions: gelatin_factory

·
Premium Member
Joined
·
5,180 Posts
Quote:
Originally Posted by Seronx View Post

They were video games...rofl
Really now...even when you add the unchanged to the i7 2700K, the FX8150 still wins
System C (Intel Core i7-2700K): 40 Votes, 40+28 = 68
System D (AMD FX-8150): 73 Votes
No Difference: 28 Votes
As a statistician, I do a lot of statistical analysis. So I have to point out when a test is flawed (or biased). Because the test is sponsored by AMD, it leaves uncertainty in the test.

Take it this way, if Intel conducted a blind test with a sample size of 200 people and concluded that 2700K blew FX-8150 out of water, would it be believable in any way?
 

·
Registered
Joined
·
636 Posts
But seriously, for the results to be considered significant AMD should have been blinded to which setup was which (more difficult to do but not impossible by any means.) Otherwise the experiment is subject to bias, unintentional or not.
 

·
Registered
Joined
·
3,015 Posts
Quote:
Originally Posted by Seronx View Post

They were video games...rofl
Really now...even when you add the unchanged to the i7 2700K, the FX8150 still wins
System C (Intel Core i7-2700K): 40 Votes, 40+28 = 68
System D (AMD FX-8150): 73 Votes
No Difference: 28 Votes
Stop ignoring the fact it was an AMD sponsored test. If Intel released something like this, YOU would be saying the exact same thing.

OT: So.. we finally got 5 core BDs? Yay, they're catching up. (And don't send me messages about how they're really (8-10) cores, look at AMD's patent. Module means core. Find it. Examine it. Learn it.)
 

·
Frequency is Megabytes
Joined
·
2,997 Posts
Quote:
Originally Posted by gelatin_factory View Post

But seriously, for the results to be considered significant AMD should have been blinded to which setup was which (more difficult to do but not impossible by any means.) Otherwise the experiment is subject to bias, unintentional or not.
They did it was a blind test...they didn't know till the end read it but I wasn't there just reading the article but since the Legit Review guys were there I have to take their word on what happened
Quote:
Originally Posted by pursuinginsanity View Post

Stop ignoring the fact it was an AMD sponsored test. If Intel released something like this, YOU would be saying the exact same thing.
I have not seen one news article where AMD has beaten Intel's legs

Intel doesn't let their Compiler to optimize AMD CPUs and doesn't adopt AMD instructions unless it is for survival
AMD does optimize their Compiler to optimize Intel CPUs and does adopt Intel instructions regardless of survival

Quote:
Originally Posted by jrbroad77 View Post

Spoiler alert: the 20% performance gain in Vishera is due to 25% more cores. You guys can do the math
thumb.gif
The performance gain from a single core is about 2x depending on what the benchmark tests for
 
1 - 20 of 136 Posts
Top