eddieroolz
Splendid
- Sep 6, 2008
- 7,057
- 0
- 25,860
- Status
blazerthon and justposting23: You both are completely wrong.
Benchmarks are up on Engadget, a 1.5ghz dual core Krait destroys 2.0ghz Medfield.
Any alleged IPC advantage for Medfield is greatly exaggerated; The previous conclusion was reached when a single core Atom barely tied a quad-core Krait in a single-threaded benchmark while consuming more power. The conclusion: OMG, ARM was almost as fast, but needed 4 cores to match Intel's 1 core!!!!!
Which is basically the same way that Intel got "twice as fast as AMD", even though the actual difference(according to Tom's) is 14% for $250 CPU vs. $250 CPU, and yes, most of those benchmarks weren't very multithreaded either. Hardly double, and hardly noticeable for that matter.
Benchmarks are up on Engadget, a 1.5ghz dual core Krait destroys 2.0ghz Medfield.
Any alleged IPC advantage for Medfield is greatly exaggerated; The previous conclusion was reached when a single core Atom barely tied a quad-core Krait in a single-threaded benchmark while consuming more power. The conclusion: OMG, ARM was almost as fast, but needed 4 cores to match Intel's 1 core!!!!!
Which is basically the same way that Intel got "twice as fast as AMD", even though the actual difference(according to Tom's) is 14% for $250 CPU vs. $250 CPU, and yes, most of those benchmarks weren't very multithreaded either. Hardly double, and hardly noticeable for that matter.
blazorthon
Glorious
- Sep 24, 2010
- 14,022
- 21
- 43,265
[citation][nom]enter_the_matrix[/nom]blazerthon and justposting23: You both are completely wrong.Benchmarks are up on Engadget, a 1.5ghz dual core Krait destroys 2.0ghz Medfield.Any alleged IPC advantage for Medfield is greatly exaggerated; The previous conclusion was reached when a single core Atom barely tied a quad-core Krait in a single-threaded benchmark while consuming more power. The conclusion: OMG, ARM was almost as fast, but needed 4 cores to match Intel's 1 core!!!!!Which is basically the same way that Intel got "twice as fast as AMD", even though the actual difference(according to Tom's) is 14% for $250 CPU vs. $250 CPU, and yes, most of those benchmarks weren't very multithreaded either. Hardly double, and hardly noticeable for that matter.[/citation]
A 1.5GHz dual-core Krait CPU does not destroy it according to Endgadget's benchmarks:
http://www.engadget.com/2012/09/18/motorolas-razr-i-benchmarks-intel-2ghz-medfield/
This is a dual core versus a single core. Excluding the CF-bench, the Medfield CPU clearly has far greater performance per core and per core per Hz. Furthermore, performance per core Hz and IPC are not the same.
Furthermore, I'm quite confused as to what you're going on about with the AMD versus Intel comparison. At stock core configuration, Sandy Bridge has a roughly 50% lead in integer performance per Hz per core over Bulldozer (although the i3s and below might not quite keep up to that as well as the i5s and i7s because of the smaller amounts of cache).
Playing any game that relies heavily one one or two threads lets Sandy Bridge easily come out far ahead of comparable Bulldozer CPUs with both at stock and mere CPU overclocking on the i5s against any Bulldozer CPU isn't enough to solve that, you have to do more complex stuff such as CPU/NB frequency overclocking and altering the core configuration to really comete with Sandy and Ivy Bridge.
Furthermore, performance per core per Hz (or IPC as you incorrectly referred to it) based benchmarks don't care about core count. If an Intel CPU can beat another quad core CPU in performance per core at a lower frequency, then it still has greater performance per Hz per core regardless of the core count differences.
A 1.5GHz dual-core Krait CPU does not destroy it according to Endgadget's benchmarks:
http://www.engadget.com/2012/09/18/motorolas-razr-i-benchmarks-intel-2ghz-medfield/
This is a dual core versus a single core. Excluding the CF-bench, the Medfield CPU clearly has far greater performance per core and per core per Hz. Furthermore, performance per core Hz and IPC are not the same.
Furthermore, I'm quite confused as to what you're going on about with the AMD versus Intel comparison. At stock core configuration, Sandy Bridge has a roughly 50% lead in integer performance per Hz per core over Bulldozer (although the i3s and below might not quite keep up to that as well as the i5s and i7s because of the smaller amounts of cache).
Playing any game that relies heavily one one or two threads lets Sandy Bridge easily come out far ahead of comparable Bulldozer CPUs with both at stock and mere CPU overclocking on the i5s against any Bulldozer CPU isn't enough to solve that, you have to do more complex stuff such as CPU/NB frequency overclocking and altering the core configuration to really comete with Sandy and Ivy Bridge.
Furthermore, performance per core per Hz (or IPC as you incorrectly referred to it) based benchmarks don't care about core count. If an Intel CPU can beat another quad core CPU in performance per core at a lower frequency, then it still has greater performance per Hz per core regardless of the core count differences.
[citation][nom]subasteve5800[/nom]2 GHz? Sounds like they are using clock speed to make up for an IPC deficit. Doesn't bode well for battery life.[/citation]
Ah, at last, someone with some genuine insight! It's Intel at its old "Megahertz Wars" game all over again. It sees a subculture of the smartphone market as the "feature counters" that eschew everything but the most storage and the highest clock rates for CPUs and hoping they will raise Atom-based phones up and trash everything else because nothing has the clock rate lead.
May Intel learn the hard way that a chip by itself does not a smartphone make.
Ah, at last, someone with some genuine insight! It's Intel at its old "Megahertz Wars" game all over again. It sees a subculture of the smartphone market as the "feature counters" that eschew everything but the most storage and the highest clock rates for CPUs and hoping they will raise Atom-based phones up and trash everything else because nothing has the clock rate lead.
May Intel learn the hard way that a chip by itself does not a smartphone make.
blazorthon
Glorious
- Sep 24, 2010
- 14,022
- 21
- 43,265
[citation][nom]gfair[/nom]Ah, at last, someone with some genuine insight! It's Intel at its old "Megahertz Wars" game all over again. It sees a subculture of the smartphone market as the "feature counters" that eschew everything but the most storage and the highest clock rates for CPUs and hoping they will raise Atom-based phones up and trash everything else because nothing has the clock rate lead.May Intel learn the hard way that a chip by itself does not a smartphone make.[/citation]
It's clocked a mere 25% higher than the previous Medfield CPU and it has greater performance per core per Hz and IPC than any ARM CPU except maybe Apple's A6. Having a high frequency doesn't mean that it much have low IPC and low performance per Hz. It's a single-core CPU, so 2GHz is not unreasonable at all.
It's clocked a mere 25% higher than the previous Medfield CPU and it has greater performance per core per Hz and IPC than any ARM CPU except maybe Apple's A6. Having a high frequency doesn't mean that it much have low IPC and low performance per Hz. It's a single-core CPU, so 2GHz is not unreasonable at all.
[citation][nom]blazorthon[/nom]Maybe, but maybe not. It wouldn't need to take any performance hit and we've already seen performance of the older Medfield CPUs and it was quite good. If necessary, the powerful FPU could be used to negate any performance impact by running emulation/translation on the FPU, but that'd also be assuming that Intel doesn't simply add other hardware acceleration for it. As for losing compatibility with native x86 programming, that's unlikely and even if it is true, it shouldn't be a big deal for CPUs that are intended to run Android.[/citation]
Yeah, I agree, that it is not impossible, although I don't think it is likely that they have any compatibility, but that is why I asked in the first place. =)
Yeah, I agree, that it is not impossible, although I don't think it is likely that they have any compatibility, but that is why I asked in the first place. =)
face-plants
Distinguished
- Apr 17, 2009
- 161
- 5
- 18,685
Some more info on the Razr i (Intel Edt) not mentioned:
-The cpu will be a single core but does support hyper-threading. This partially answers many people's questions as to why they've ignored the popular dual- and quad-core smartphone chip trend.
-The camera on the Intel version can be ready to snap a picture in under a second from the home screen (twice as fast as the best competitor so they say). It also offers a high speed mode which takes 10 pictures per second which is faster than many DSLRs out there. Both of these features are thanks to a new image signal processor Intel specifically designed into their chipset to work with Motorola's technology.
-Also, the phone will lack the Chrome browser pre-loaded (although it is still available as a free download from the Play store). The reason given for this was that Chrome has hardware-accelerated page rendering on ARM chips. (I'm guessing it doesn't run so great on the intel chipset yet). Chrome will be preloaded once they have "complete chipset optimization" for it. Weird choice but I guess they don't want any apps on it out of the box that dont run perfectly.
All in all, it's a decent showing from Intel as they try to break into the mobile market. They're not breaking any records (except for picture speed maybe), but it's a solid entry. At the very least, Intel will keep the other mobile chip designers on their toes and keep things competitive.
-The cpu will be a single core but does support hyper-threading. This partially answers many people's questions as to why they've ignored the popular dual- and quad-core smartphone chip trend.
-The camera on the Intel version can be ready to snap a picture in under a second from the home screen (twice as fast as the best competitor so they say). It also offers a high speed mode which takes 10 pictures per second which is faster than many DSLRs out there. Both of these features are thanks to a new image signal processor Intel specifically designed into their chipset to work with Motorola's technology.
-Also, the phone will lack the Chrome browser pre-loaded (although it is still available as a free download from the Play store). The reason given for this was that Chrome has hardware-accelerated page rendering on ARM chips. (I'm guessing it doesn't run so great on the intel chipset yet). Chrome will be preloaded once they have "complete chipset optimization" for it. Weird choice but I guess they don't want any apps on it out of the box that dont run perfectly.
All in all, it's a decent showing from Intel as they try to break into the mobile market. They're not breaking any records (except for picture speed maybe), but it's a solid entry. At the very least, Intel will keep the other mobile chip designers on their toes and keep things competitive.
ojas
Distinguished
- Feb 25, 2011
- 2,924
- 0
- 20,810
Well, this is great, but i think they'll need a much larger battery than the one they put inside the Xolo X900 if they're bumping the max clock speed by 400 MHz.
They do have power management features enabled, so there are various idle and active states.
Android app compatibility: Google Play won't let you download apps that aren't supported. The SoC makes use of an ARM translation layer so most of them can be used, some like temple run, Chrome, Opera (opera mini works), firefox, and a few others didn't work till late july at least.
Don't know what the situation will be like with ICS. But i think power consumption and compatibility would have improved. It lasts at least 6 hours with intense multimedia, and over a day with normal phone-related use.
The one i reviewed had a battery related bug, but i don't know how wide-spread it is or if it's a hardware issue or a software issue.
Also, no, these Atoms don't do GFLOPS, only MFLOPS (and it should be MIPS in the chart, not GIPS):
They do have power management features enabled, so there are various idle and active states.
Android app compatibility: Google Play won't let you download apps that aren't supported. The SoC makes use of an ARM translation layer so most of them can be used, some like temple run, Chrome, Opera (opera mini works), firefox, and a few others didn't work till late july at least.
Don't know what the situation will be like with ICS. But i think power consumption and compatibility would have improved. It lasts at least 6 hours with intense multimedia, and over a day with normal phone-related use.
The one i reviewed had a battery related bug, but i don't know how wide-spread it is or if it's a hardware issue or a software issue.
Also, no, these Atoms don't do GFLOPS, only MFLOPS (and it should be MIPS in the chart, not GIPS):
- Status
TRENDING THREADS
-
News First credible report of RTX 5090 FE with melted connector appears — third-party cable likely cause- Started by Admin
- Replies: 52
-
-
-
Discussion What's your favourite video game you've been playing?- Started by amdfangirl
- Replies: 4K
Facebook
Twitter
Instagram