Apollo Lake vs Braswell Geekbench Comparison
Highest Pentium N3710 (4C/4T up to 2.56 GHz) Windows x86 Score:
Single-Core Score: 1030
Multi-Core Score: 3541
https://browser.primatelabs.com/geekbench3/5378070
Highest Pentium N4200 (4C/4T up to 2.4/2.5 GHz) Windows x86 Score:
Single-Core Score: 1418
Multi-Core Score: 4601
https://browser.primatelabs.com/geekbench3/7251504
Goldmont cores should bring a nice boost to ST performance, great news for budget notebooks.
Is the Goldmont architecture going to show up in cheap tablets? It would be a great boost for that segment, but IMO they are still under-powered for laptops.
But not under-powered for cheap laptops. Although they may be little more than tablets with attached keyboards, the large screens and stiff hinges of cheap laptops give them a better form factor. For us old guys in particular, those of us who need reading glasses, we need a larger screen in a package light enough to carry. Cheap laptops do the trick. Plus, they're cheap.
Apollo Lake vs Braswell Geekbench Comparison
Highest Pentium N3710 (4C/4T up to 2.56 GHz) Windows x86 Score:
Single-Core Score: 1030
Multi-Core Score: 3541
Highest Pentium N4200 (4C/4T up to 2.4/2.5 GHz) Windows x86 Score:
Single-Core Score: 1418
Multi-Core Score: 4601
Goldmont cores should bring a nice boost to ST performance, great news for budget notebooks.
The ST performance increase in integer when putting AES and other crypto stuff is only about 20%.
OTOH FP got almost 50% better, and memory got 30% better.
Is the Goldmont architecture going to show up in cheap tablets? It would be a great boost for that segment, but IMO they are still under-powered for laptops.
The more interesting comparison is with an N3700 because only 5% of Braswell based Nxxx devices are shipped with an N3710, the other 95% are shipped with N3700.
This is compared to the fastest x86 N3700 result:
https://browser.primatelabs.com/geekbench3/compare/4289803?baseline=7251504
28% faster Integer Singlethread and 25% faster Integer Multithread if we exclude all the AES and SHA scores. Floating Point Performance is over 50% better on Goldmont. Maybe Intel was more focused on Floating point improvements because the Floating Point IPC disadvantage was much worse compared to be big cores like Skylake.
How does it compare to Jaguar/Puma?
Almost core 2 performance. Not bad for a chip like that.
Indeed. Single-threaded speed has been the major drawback of Atom-based PCs.No its faster than core 2 quad performance. A core 2 quad (q6600 at 2.4ghz) gets about 1200 to 1300 in geekbench for the single core vs this pentium apollo lake atom which rumored max clock speed is 2.5 ghz and it is getting in the upper 1300 to low 1400 in geekbench single thread.
Clock for clock it looks like the new apollo lake atom is the same speed if not faster than core 2 quad of the 65nm or 45nm generation. This is a very big deal.
I did not know that. Intel "promising" Phenom II X6 1100T performance, in an Atom CPU? That would be crazy. Maybe another shrink and they'll get there?That said it is still disappointing for in 2011 intel was promising AMD 1100t x6 phenom ii performance with goldmount/ apollo lake. Apollo Lake is simply just not there, but if they had another 30% clock speed and 2 more cores it would be there trading blows with a stock 1100t x6 phenom ii (pretty much the best of the phenom ii generation). Now why did Intel pick the phenom 1100t as a comparison? Well that was the best of AMD at the time.
That said it is still disappointing for in 2011 intel was promising AMD 1100t x6 phenom ii performance with goldmount/ apollo lake.
I remember that they said 45nm Core 2 performance with Goldmont, however they only managed to get 65nm.Do you have a link to this promise?
Will Goldmont beat my Pentium M score in CPUmark99? If not, I'll stick with my 2004 laptop CPU haha.No its faster than core 2 quad performance. A core 2 quad (q6600 at 2.4ghz) gets about 1200 to 1300 in geekbench for the single core vs this pentium apollo lake atom which rumored max clock speed is 2.5 ghz and it is getting in the upper 1300 to low 1400 in geekbench single thread.
Only when you set the baseline of core 2 quad at the 3.0 ghz 12mb cache q9650 or an overclocked 3 to 3.6 ghz q6600 will you see this number as "under" core 2 performance.
Clock for clock it looks like the new apollo lake atom is the same speed if not faster than core 2 quad of the 65nm or 45nm generation. This is a very big deal.
-----
This is great news, for hopefully storage speed whether emmc or ssd or hard drive is loads better than the crappy hard drives we had in 2006. Plus the integrated graphics of these apollo lake atoms are going to be loads better than the integrated graphics of 2006, and even better than most of the graphic cards in 2006 until you spend north of $200 for a graphic card that had a tdp of around 120 to 150watts (still a single 75w 6 pin for additional board power but still getting as close to that 150w as possible).
In other words as long as apollo lake devices are at least 4 cores, at least 2gb of memory (but hopefully windows will move the OEM licensing free cost to 4gbs), and a decent storage the bare minimum hardware sold to run windows or android will have its baseline moved so it will be less sucky for everyone. We can finally now run decent office and decent web apps and video on hardware that is so cheap we call it "crap hardware" or "crap tops."
Put another way this hardware shits all over the 1st gen macbook air (2008), and even is better than the 2nd generation macbook air (2010). It is not until 2011 where we seen the 2nd gen core i series that you will see the macbook airs beating this hardware and even then due to lower clock speeds using turbo to get better clcok speeds and the fact we are talking 2 core / 4 thread devices the 2011 macbook airs will be trading blows with the airs being faster in single thread but slower in multi thread. The 2nd generation macbook airs are also the same time Intel released its ultrabook spec.
So as long as we get good enough ram and good enough storage we are talking crap laptops that are competing against 2011 ultra portable hardware, or we are talking 2008 to 2009 desktop hardware but now in a 6w tdp and thus you will see lots of 10 to 13 inch computers that are ultraportable but also lots of 14 to 15" cheap laptops.
-----
That said it is still disappointing for in 2011 intel was promising AMD 1100t x6 phenom ii performance with goldmount/ apollo lake. Apollo Lake is simply just not there, but if they had another 30% clock speed and 2 more cores it would be there trading blows with a stock 1100t x6 phenom ii (pretty much the best of the phenom ii generation). Now why did Intel pick the phenom 1100t as a comparison? Well that was the best of AMD at the time.
So did Intel intentionally lie and overpromise and under deliver in their 2011 marketing to investors?
Not neccessarly at the time Intel had far more hopes for its 14nm process than what actually came out years later.
If Intel was targeting a 3.0 to 3.3 ghz ghz clock speed 5 years ago (instead of a 2.5 ghz clock speed) before they knew all of the 14nm characteristics than this number could have been possible and literally on the nose of the similar AMD hardware in Intel's mind.
Do you have a link to this promise?
Indeed. Single-threaded speed has been the major drawback of Atom-based PCs.
I did not know that. Intel "promising" Phenom II X6 1100T performance, in an Atom CPU? That would be crazy. Maybe another shrink and they'll get there?
No its faster than core 2 quad performance. A core 2 quad (q6600 at 2.4ghz) gets about 1200 to 1300 in geekbench for the single core vs this pentium apollo lake atom which rumored max clock speed is 2.5 ghz and it is getting in the upper 1300 to low 1400 in geekbench single thread.
Only when you set the baseline of core 2 quad at the 3.0 ghz 12mb cache q9650 or an overclocked 3 to 3.6 ghz q6600 will you see this number as "under" core 2 performance.
Clock for clock it looks like the new apollo lake atom is the same speed if not faster than core 2 quad of the 65nm or 45nm generation. This is a very big deal.
.
Note this is not the only source, but the 1st one google polled up. It was talking about some marketing slides from an intel developer conference.
http://www.geek.com/chips/14nm-atom...6-core-amd-phenom-ii-10x-todays-atom-1407293/
Also to be perfectly fair to intel Airmont is canceled
Airmont was going to be the lower tdp version of Goldmont CPUs and Airmont was meant for tablets, while Apollo Lake is the higher tdp version meant for ultra convertible tablets and larger form factors.
As some of you remember, we did a short introduction, spiced up with more inside info, of the upcoming budget-centric Apollo Lake processors from Intel but now we are glad to tell you that a pre-production sample with Pentium N4200 just arrived in our office.
Aside from the significant energy consumption and performance increase, the CPU should bring support to some much-needed features in the low-end segment of the market and we will take the time to test it out in the following days. For now, we will let you know that the Pentium N4200 sports 4 cores clocked at 1.1 GHz and can go way up to 2.5 GHz. While the base frequency is much lower than the N3700 from the Braswell generation we reviewed a while back, the maximum burst frequency is slightly higher. Either this is an early sample unit or Intel has decided to lower the base frequency in exchange for prolonged battery life. Still, the TDP is ~6W and the cTDP is ~4W just like before.
Goldmont cores should bring a nice boost to ST performance, great news for budget notebooks.
LaptopMedia is testing a Pentium N4200 (B0 ES).
One of the first Apollo Lake chips, the Intel Pentium N4200, surfaces with detailed specs
http://laptopmedia.com/news/one-of-...el-pentium-n4200-surfaces-with-detailed-specs
This Christmas I would like to get my mother-in-law a new laptop to replace her old single-core Atom-based system. Apollo Lake should be perfect for this.