The Intel Atom Thread

Page 210 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
That's an improvement of 12.5%, that's not that much (especially considering the skipped Gen 10).

That's incorrect reasoning.

You got 2x gains going from 24EU Gen 9 to 64EU Gen 11, while the bandwidth increased 55%. That's because things don't scale linearly.

Bandwidth increases only by 33% while the shader throughput increases by 77% for Elkhart Lake.

If you want double the performance in a balanced uarch, you need to double the bandwidth, the fillrate, and the shader throughput.

You agree yea?
 
Last edited:
  • Like
Reactions: mikk

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
It means Gen11 is making better use of available bandwidth.

Something like that.

Prescott Pentium 4 was not faster than Northwood.

Yet on Celerons, it was 25-30% faster per clock.

Same thing with current CPUs where Core class chips are seeing 15-20% gains but Atom-class chips are getting 30%.
 

moinmoin

Diamond Member
Jun 1, 2017
4,926
7,608
136
That's nice and all. My point was that whereas the platform balancing optimizations are all commendable the architecture improvements appear to be negligible. That the Cores are even worse doesn't make it any better.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
That's nice and all. My point was that whereas the platform balancing optimizations are all commendable the architecture improvements appear to be negligible. That the Cores are even worse doesn't make it any better.

Architectural improvements(larger caches, tiling) are why it can perform greater than specs suggest.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
Gemini Lake Refresh J5040 gets 95 points in Cinebench R15 1T. That's a 15% improvement. That's a 3-4% higher than specs suggest. It must not having been reaching boost as reliably.

Improvement will be less in MT. Seems to be 5-7%.
 

DrMrLordX

Lifer
Apr 27, 2000
21,570
10,763
136
Gemini Lake Refresh J5040 gets 95 points in Cinebench R15 1T. That's a 15% improvement. That's a 3-4% higher than specs suggest. It must not having been reaching boost as reliably.

Improvement will be less in MT. Seems to be 5-7%.

Hmm, do we have any idea what kind of average clock it's hitting in ST and MT?
 

vstar

Member
May 8, 2019
46
39
61
First 10nm Tremont SOC?

I think this is the first news of a Tremont SOC (other than Lakefield) to come out this year.

I didn't see anything else about the technical specs for the P5900, but I am curious to see what the ST/MT clock speeds are.
 
  • Like
Reactions: coercitiv

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
Looking at Lakefield die, I estimate each Tremont cores to be 0.7mm2 without the 512KB L2 cache.

So its actually comparable to ARM cores with similar uarch/performance using 7nm process.

It's only their Core cores that are very large. Speculation that it must be larger to allow for high clocks could very well be true. The two AVX-256 units in Sunny Cove are almost the size of one Tremont core along with its 512KB L2 cache.
 
Last edited:

Thala

Golden Member
Nov 12, 2014
1,355
653
136
So its actually comparable to ARM cores with similar uarch/performance using 7nm process.

It is close to impossible to design an x86-64 CPU with the same relative performance in the same power and area range compared to ARM...so i pretty much have doubt in your claim.

But yes, depending on your cell library, the high performance cells can be larger - this holds in particular for SRAM cells. In the past for instance you had different track heights. For N7 you can mostly chose different Vth cells and not different track heights.
 
Last edited:

TheGiant

Senior member
Jun 12, 2017
748
353
106
It is close to impossible to design an x86-64 CPU with the same relative performance in the same power and area range compared to ARM...so i pretty much have doubt in your claim.
relatively impossible to tech calculated output or the current and "D+1" software connected to it?
I doubt ARM can replace x86 if AMD and Intel deliver
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
Gracemont features double the L1 instruction cache - 64KB. That's probably important as it needs to feed the dual instruction stream introduced with Tremont.
 
  • Like
Reactions: Dayman1225

you2

Diamond Member
Apr 2, 2002
5,668
913
126
Given AMD and snapdragon progress is there any reason to actually consider atom ?
 

you2

Diamond Member
Apr 2, 2002
5,668
913
126
This thread started in 2013 before you joined. Times change. It was a question - why are you avoiding the question ?
 
  • Like
Reactions: AkulaMD

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
This thread started in 2013 before you joined. Times change. It was a question - why are you avoiding the question ?

Err, don't you have all the information to consider whether its suitable for you then?

It's quite simple. For example if you need low power x86 you probably do. If not, an ARM core.
 

you2

Diamond Member
Apr 2, 2002
5,668
913
126
My point was technology has changed in 7 years and i was asking if atom was still relevant given new offerings that didn't exist 2013. It was simple enough. I've not tracked amd and snapdragon close enough to know how they compare with regards to cost, power and performance (though i suspect snapdragon beats intel hand downs on performance - esp graphical). I do wonder if there is an x86 emulator that will run on snapdragon - i know for example IBM mostly runs emulator on their risc processors for many of their newer mainframes (crise they must make a killing on those machines).

Err, don't you have all the information to consider whether its suitable for you then?

It's quite simple. For example if you need low power x86 you probably do. If not, an ARM core.
 

Roland00Address

Platinum Member
Dec 17, 2008
2,196
260
126
My point was technology has changed in 7 years and i was asking if atom was still relevant given new offerings that didn't exist 2013. It was simple enough. I've not tracked amd and snapdragon close enough to know how they compare with regards to cost, power and performance (though i suspect snapdragon beats intel hand downs on performance - esp graphical). I do wonder if there is an x86 emulator that will run on snapdragon - i know for example IBM mostly runs emulator on their risc processors for many of their newer mainframes (crise they must make a killing on those machines).
This thread is not single about a single atom generation, hell it is not even about all the things branded "atom" instead it is about intel's small core which has purposes besides being cheaper for intel to produce.
So yeah this 200 page thread is not about a single atom but is tracking atom over time. Furthermore atom the small core is not a single thing, having a 2.5w smartphone soc is different than a tablet, which is different than a laptop / netbook, which is different than embedded (think a nas server), than a server which is trying to do very low powerful server stuff that wants cores but does not care if its fast cores just low power energy efficient cores.

And yes there are comparisons against big core intel, arm generic, and specifically qualcomm and things like windows on arm (which is qualcomm only so far.)
 
  • Love
Reactions: spursindonesia

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
My point was technology has changed in 7 years and i was asking if atom was still relevant given new offerings that didn't exist 2013.

Emulation by nature has flaws and cannot ever be anything more than a stepping stone. It's not just about loss in performance, there's often significant loss in compatibility that articles do not discuss. Of course, its trying to be something its not.

So ultimately it still depends on you. You could have asked a technical question but you did not.
 

you2

Diamond Member
Apr 2, 2002
5,668
913
126
On this i will disagree because in a sense even the processors are running 'emulation' through the microcode installed on the hardware. The key difference with emulation on another chip would (imho) be timing. The performance of instructions will be different. Now it is true that a user (or even kernal mode) emulator might have other differences since the hardware rules might fundamentally be different (a simple example might be the page size) still I believe you can get emulation so that better than 99.99% of the software will run correctly (ignoring performance impact). You might uncover bugs in the software that might otherwise not shown up until moved to a later generation of intel code (for example software that make assumption around current implementation of branch prediction or instruction timings since x86 does not promise these behavior will remain the same across processor generation - but of course some of these flaws show up when you move from intel to amd or even within the same vendor processor (atom to desktop processor).
-

Emulation by nature has flaws and cannot ever be anything more than a stepping stone. It's not just about loss in performance, there's often significant loss in compatibility that articles do not discuss. Of course, its trying to be something its not.

So ultimately it still depends on you. You could have asked a technical question but you did not.