Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Haswell CPU could be 10% faster

dusk007

dusk007

macrumors 68040
Dec 5, 2009
3,411
104
Ploki said:
Actually revolution = not necessarily better, just new.
Click to expand...
And what did I write?
dusk said:
Revolution is something entirely new. It is a big sudden change.
Click to expand...
Ploki said:
Evolution doesn't necessary have to be slow you know? It just happens that evolution of life was slow.
Click to expand...
Not sure what you are reading but since you quote me, I didn't write slow. I wrote gradual and building on the old. Not turn around new. Not back to the drawing board new.
dusk said:
Evolution is anything that is better than the previous however little.
Click to expand...
Ploki said:
Graphene transistors would be in fact evolution. Except if you would change everything you know about a computer...
Click to expand...
Based on what the folks say they can do it would be a revolution in the transistor department. It is always a matter of perspective.
From the point of view of a PC there is never going to be anything revolutionary in the tech. Only the use of it can be. From the tech's perspective you may at some point do something entirely different like run QM algorithms. Which would be quite a big thing in cryptography. You could throw a few things away and could use some really cool new ones.

revolution means turn around.
evolution meas to gradually roll out or unfold.
The words aren't related in any significant way despite there being only one letter in difference.

bill-p simply used the words wrong.
 
B

bill-p

macrumors 68030
Jul 23, 2011
2,889
1,550
dusk007 said:
No you do have it backwards. Evolution may lead to significant change at some point but it is gradual. It builds upon the old and only changes a little each generation.
Revolution is chop of everybody's head we are doing it completely different than it used to be. Revolution is something entirely new. It is a big sudden change.

Evolution is anything that is better than the previous however little. Revolution would be if they pull graphene transistors or something actually new out of their butts. Diamond chips or something that would be easier to cool.
Click to expand...

Look at it this way...

Haswell: Intel takes the same base concept (Ivy Bridge), increase GPU performance instead of CPU performance, integrate more of the chipsets. Sounds like the "opposite" of what they'd usually do (increase CPU performance). In which case, it's a "revolution". Like you said, it's "different from how it used to be".

Broadwell: Intel moves to a new manufacturing process (14nm) and increase CPU performance, which is what they usually do. Sounds like an "evolution" to me.

----

But let's disregard the terminology, and go back to the discussion. The problem here is that I'm sure some people are expecting Haswell to be this magical improvement over Ivy Bridge where they can have all of the cake: faster CPU, faster GPU, lower power consumption.

But the reality is that it may not be all that. And we got all of this information from last year. But I guess... people just somehow talk themselves into believing that technological marvels come every year... thanks to what Apple did to the iOS platform.
 
P

paulrbeers

macrumors 68040
Dec 17, 2009
3,963
123
Ploki said:
Actually revolution = not necessarily better, just new.
Evolution doesn't necessary have to be slow you know? It just happens that evolution of life was slow.
Graphene transistors would be in fact evolution. Except if you would change everything you know about a computer...
Click to expand...

I'm thinking you don't understand the meaning of revolutionary....

Merriam-Webster devices Revolutionary as: constituting or bringing about a major or fundamental change

Evolution: a process of continuous change from a lower, simpler, or worse to a higher, more complex, or better state

There for Dusk007 is right that it should have been Haswell is evolutionary and not revolutionary.
 
P

Ploki

macrumors 601
Jan 21, 2008
4,308
1,558
paulrbeers said:
I'm thinking you don't understand the meaning of revolutionary....

Merriam-Webster devices Revolutionary as: constituting or bringing about a major or fundamental change

Evolution: a process of continuous change from a lower, simpler, or worse to a higher, more complex, or better state

There for Dusk007 is right that it should have been Haswell is evolutionary and not revolutionary.
Click to expand...

New chip material isn't something that's going to revolutionise the personal computer for the end user... Perspective
 
P

paulrbeers

macrumors 68040
Dec 17, 2009
3,963
123
bill-p said:
Look at it this way...

Haswell: Intel takes the same base concept (Ivy Bridge), increase GPU performance instead of CPU performance, integrate more of the chipsets. Sounds like the "opposite" of what they'd usually do (increase CPU performance). In which case, it's a "revolution". Like you said, it's "different from how it used to be".
.
Click to expand...

Yeah, but that's not the context you used, to quote you:

I have said this a while before, but Haswell is not a magical fix for whatever performance problem Ivy Bridge has now. It's a revolutionary cycle, not an evolutionary cycle.

In the context of that sentence, since Haswell is not a magical fix for what ails Ivy Bridge, it can only be considered an evolutionary change since all Intel is doing is upping the processes completed per cycle and increasing the GPU. To be revolutionary, haswell would have to be a completely different architecture that either blows the old architecture out of the water OR comes from a completely different thought process (i.e. Pentium 4 to Core2duo which did both with lower clocks but faster architecture).

Edit: Also remember, this is EXACTLY what they did from Sandy Bridge to Ivy Bridge -> CPU only somewhat faster, but GPU 2x as fast. So again this isn't even a "revolutionary" approach to CPU/GPU's for Intel.

----------
Ploki said:
New chip material isn't something that's going to revolutionise the personal computer for the end user... Perspective
Click to expand...

Thanks for making my point! :) So Bill-P had it wrong and Duck007 was right that Bill-P was wrong.... Haswell is evolutionary and not revolutionary.
 
B

bill-p

macrumors 68030
Jul 23, 2011
2,889
1,550
paulrbeers said:
Yeah, but that's not the context you used, to quote you:

I have said this a while before, but Haswell is not a magical fix for whatever performance problem Ivy Bridge has now. It's a revolutionary cycle, not an evolutionary cycle.

In the context of that sentence, since Haswell is not a magical fix for what ails Ivy Bridge, it can only be considered an evolutionary change since all Intel is doing is upping the processes completed per cycle and increasing the GPU. To be revolutionary, haswell would have to be a completely different architecture that either blows the old architecture out of the water OR comes from a completely different thought process (i.e. Pentium 4 to Core2duo which did both with lower clocks but faster architecture).

Edit: Also remember, this is EXACTLY what they did from Sandy Bridge to Ivy Bridge -> CPU only somewhat faster, but GPU 2x as fast. So again this isn't even a "revolutionary" approach to CPU/GPU's for Intel.
Click to expand...


Oh dear... and then we keep going back on this. Let's agree to disagree then.

On that note, this NOT what Intel did from Sandy Bridge to Ivy Bridge. Except for synthetic benchmarks, Ivy Bridge GPU was only on average about 50% faster than Sandy at most. Not 2x.

Also Ivy on the CPU side is typically 20% faster than Sandy, and power consumption has dropped by a lot as well.
http://www.anandtech.com/show/5626/ivy-bridge-preview-core-i7-3770k/10

That's a lot more than what Haswell is doing with Ivy (only 10% improvement, no power consumption improvement).

And in case you missed it, Intel also moved to 3D tri-gate transistors with Ivy.
http://www.anandtech.com/show/4313/...nm-3d-trigate-transistors-shipping-in-2h-2011

Haswell isn't moving to any new tech as compared to Ivy Bridge.
 
P

paulrbeers

macrumors 68040
Dec 17, 2009
3,963
123
bill-p said:
Oh dear... and then we keep going back on this. Let's agree to disagree then.

On that note, this NOT what Intel did from Sandy Bridge to Ivy Bridge. Except for synthetic benchmarks, Ivy Bridge GPU was only on average about 50% faster than Sandy at most. Not 2x.

Also Ivy on the CPU side is typically 20% faster than Sandy, and power consumption has dropped by a lot as well.
http://www.anandtech.com/show/5626/ivy-bridge-preview-core-i7-3770k/10

That's a lot more than what Haswell is doing with Ivy (only 10% improvement, no power consumption improvement).

And in case you missed it, Intel also moved to 3D tri-gate transistors with Ivy.
http://www.anandtech.com/show/4313/...nm-3d-trigate-transistors-shipping-in-2h-2011

Haswell isn't moving to any new tech as compared to Ivy Bridge.
Click to expand...

I didn't miss any of that. So you are saying that 20% is dramatically different than 10% increase? I'd call that evolutionary, since neither 10 nor 20 is exactly "blowing the doors off" of the CPU world now is it?

As for the difference between HD3000 -> HD4000 -> HD5600.... It's all relative. Frankly in some benchmarks the HD4000 is almost 2x, but in others there is limited increase. It really doesn't matter. For all we know the HD5600 could end up only be 50% faster. Kind of hard to say since NO ONE outside of Intel has even see/used it right?

Further, weren't the biggest gains from Sandy Bridge to Ivy Bridge then in the GPU? Since the CPU only gained 20%, but the GPU gained 50% (based on your numbers). Isn't that what they are doing (roughly) with Ivy Bridge to Haswell (GPU is going to out gain the CPU). So gain its the SAME approach.

Lastly to your point about Tri-gate, that was a revolutionary step for their manufacturing process. It didn't revolutionize the processor. If it was something magical they added to ONLY Ivy Bridge then sure, but frankly they are rolling out tri-gates to all their transistor based product. Even then crappy Atom processor is getting it. They can make everything smaller and therefore cheaper and use less energy.
 
B

bill-p

macrumors 68030
Jul 23, 2011
2,889
1,550
paulrbeers said:
I didn't miss any of that. So you are saying that 20% is dramatically different than 10% increase? I'd call that evolutionary, since neither 10 nor 20 is exactly "blowing the doors off" of the CPU world now is it?

As for the difference between HD3000 -> HD4000 -> HD5600.... It's all relative. Frankly in some benchmarks the HD4000 is almost 2x, but in others there is limited increase. It really doesn't matter. For all we know the HD5600 could end up only be 50% faster. Kind of hard to say since NO ONE outside of Intel has even see/used it right?

Further, weren't the biggest gains from Sandy Bridge to Ivy Bridge then in the GPU? Since the CPU only gained 20%, but the GPU gained 50% (based on your numbers). Isn't that what they are doing (roughly) with Ivy Bridge to Haswell (GPU is going to out gain the CPU). So gain its the SAME approach.

Lastly to your point about Tri-gate, that was a revolutionary step for their manufacturing process. It didn't revolutionize the processor. If it was something magical they added to ONLY Ivy Bridge then sure, but frankly they are rolling out tri-gates to all their transistor based product. Even then crappy Atom processor is getting it. They can make everything smaller and therefore cheaper and use less energy.
Click to expand...

Did you intentionally miss the part about power consumption also dropping with Ivy Bridge?

Haswell doesn't drop power consumption. Or at least Intel's released numbers thus far don't seem to indicate anything of the sort.

Ploki said:
this forum sometimes has the most pointless debates.
Click to expand...

Yeah. I agree. Some people get so apologetic about Intel's HD 4000 and about their CPU strategies (Haswell).

I got into the same argument over the HD 4000... which took up multiple pages for nothing at all.

I guess this is how people talked themselves into waiting for Haswell in the first place... but still, it's sort of unbelievable.
 
adjeff8

adjeff8

macrumors 6502
Nov 18, 2012
466
4
bill-p said:
Look at it this way...

Haswell: Intel takes the same base concept (Ivy Bridge), increase GPU performance instead of CPU performance, integrate more of the chipsets. Sounds like the "opposite" of what they'd usually do (increase CPU performance). In which case, it's a "revolution". Like you said, it's "different from how it used to be".

Broadwell: Intel moves to a new manufacturing process (14nm) and increase CPU performance, which is what they usually do. Sounds like an "evolution" to me.

----

But let's disregard the terminology, and go back to the discussion. The problem here is that I'm sure some people are expecting Haswell to be this magical improvement over Ivy Bridge where they can have all of the cake: faster CPU, faster GPU, lower power consumption.

But the reality is that it may not be all that. And we got all of this information from last year. But I guess... people just somehow talk themselves into believing that technological marvels come every year... thanks to what Apple did to the iOS platform.
Click to expand...

Regardless. It will be a 2nd generation 15" MacBook Pro. I've been chomping at the bit since last June and the fact that it is not 1st gen makes me feel more secure.
 
C

cirus

macrumors 6502a
Mar 15, 2011
582
0
yusukeaoki said:
Mac Pro has Xeon processors.
Click to expand...


And what does than have to do with my statement?

I never compared it to a mac (let alone a mac pro), I compared mobile processors to normal desktop processors (XENON \= i5)

Look up the cinebench scores if you don't believe me, an i7 3740m beats any non overclocked i5 desktop processor.

But anyway, any hackintosh not using an i7 or xenon, a i7 mobile will beat.

The only mac pro that has one cpu that can beat the 15 rmbp (3820m) on geekbench is the Intel Xeon W3680 3330 MHz (6 cores) one. Considering that you can get an i7 mobile laptop for less than $800 that will give you slightly less than half the performance of a 12 core dual cpu mac pro (at $5000+) I'd say mobile compares very competitively to desktop. Mobile cpus stand up very well to desktop or xenon cpus compared to mobile gpus vs desktop gpus (which are often more than order of magnitude differences). If you mean can't stand a chance to be "beaten by 50% or more" there are very few mac pros that can do that.

Anything comparing a dual cpu xenon mac pro and a i7 mobile cpu are clearly apples to oranges but anyway. Anything that can get half the performance for less than 1/5 the price clearly competes very well (I'm assuming that ecc ram, xenon is not needed because otherwise we would not be making the comparison).

For work not requiring a xenon or ECC ram, laptops can compete with desktops (especially when its not gpu related).

http://www.anandtech.com/show/6512/dell-precision-m6700-notebook-review-the-enterprise-split/4
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom