The Intel Atom Thread

Page 205 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

dark zero

Platinum Member
Jun 2, 2015
2,497
4
61
I think the point of my post was that nothing Intel does to handicap Atom (in this case, a Pentium J5005) really forces you to buy Core. No vPro and AVX2 is not exactly a big deal in a home/office AiO that is cheap.



In the same power envelope? Who in the x86 world is chasing the 5-15w space? AMD barely has anything running at such a low power.
And VIA looks away with their IPC jump...
 
Nov 29, 2008
133
1
91
Anyone testing J4005 yet?
Today I have got a board and made some tests with Linux 64bit GCC.
The classic nsieve benchmark is fast or how do you interprete this:
Double as Jaguar at 2GHz and half of Core i7-8700 at 4.4 HHz
Open MP is like 90% on the second thread.
Any J5005 i have not seen yet.

i have only 32bit tests on J1900 and nsieve is close to Jaguar.
 
Nov 29, 2008
133
1
91
JS Kraken 1.1 the same Firefox:
J4005: 2334 ms
i7-8700: 827 ms

From Anandtech (not Firefox)
J1800: 4306 ms
J1900: 4547 ms
Athlon 5350: 4555 ms
 
Nov 29, 2008
133
1
91
J4005
Sieve of Eratosthenes (Scaled to 10 Iterations)
Version 1.2b, 26 Sep 1992

Array Size Number Last Prime Linear RunTime MIPS
(Bytes) of Primes Time(sec) (Sec)
8191 1899 16381 0.000 0.000 4832.2
10000 2261 19997 0.000 0.000 4553.4
20000 4202 39989 0.001 0.001 4271.5
40000 7836 79999 0.002 0.003 3144.0
80000 14683 160001 0.003 0.005 3512.6
160000 27607 319993 0.007 0.011 2981.4
320000 52073 639997 0.013 0.024 2865.3
640000 98609 1279997 0.027 0.047 2947.8
1280000 187133 2559989 0.054 0.093 3027.5
2560000 356243 5119997 0.107 0.217 2617.0
5120000 679460 10239989 0.214 0.737 1551.8
10240000 1299068 20479999 0.429 2.358 977.5
20480000 2488465 40960001 0.858 5.447 853.2
40960000 4774994 81919993 1.716 11.934 784.7

Relative to 10 Iterations and the 8191 Array Size:
Average RunTime = 0.001 (sec)
High MIPS = 4832.2
Low MIPS = 784.7

i7-8700
Sieve of Eratosthenes (Scaled to 10 Iterations)
Version 1.2b, 26 Sep 1992

Array Size Number Last Prime Linear RunTime MIPS
(Bytes) of Primes Time(sec) (Sec)
8191 1899 16381 0.000 0.000 9241.7
10000 2261 19997 0.000 0.000 9167.5
20000 4202 39989 0.000 0.000 8856.6
40000 7836 79999 0.001 0.001 8061.8
80000 14683 160001 0.002 0.002 7782.6
160000 27607 319993 0.004 0.004 7792.3
320000 52073 639997 0.007 0.010 7064.4
640000 98609 1279997 0.014 0.022 6236.7
1280000 187133 2559989 0.028 0.047 6038.3
2560000 356243 5119997 0.056 0.094 6024.1
5120000 679460 10239989 0.112 0.189 6048.0
10240000 1299068 20479999 0.224 0.430 5358.1
20480000 2488465 40960001 0.449 1.484 3131.6
40960000 4774994 81919993 0.897 4.141 2261.5


Relative to 10 Iterations and the 8191 Array Size:
Average RunTime = 0.000 (sec)
High MIPS = 9241.7
Low MIPS = 2261.5

Athlon 5350

Sieve of Eratosthenes (Scaled to 10 Iterations)
Version 1.2b, 26 Sep 1992

Array Size Number Last Prime Linear RunTime MIPS
(Bytes) of Primes Time(sec) (Sec)
8191 1899 16381 0.001 0.001 2141.5
10000 2261 19997 0.001 0.001 2159.2
20000 4202 39989 0.002 0.002 2213.2
40000 7836 79999 0.004 0.004 2084.4
80000 14683 160001 0.008 0.009 1915.3
160000 27607 319993 0.015 0.019 1824.9
320000 52073 639997 0.030 0.039 1745.7
640000 98609 1279997 0.060 0.079 1756.7
1280000 187133 2559989 0.121 0.162 1735.7
2560000 356243 5119997 0.242 0.603 939.9
5120000 679460 10239989 0.484 2.161 529.2
10240000 1299068 20479999 0.968 5.110 451.1
20480000 2488465 40960001 1.936 11.213 414.4
40960000 4774994 81919993 3.871 24.005 390.1

Relative to 10 Iterations and the 8191 Array Size:
Average RunTime = 0.002 (sec)
High MIPS = 2213.2
Low MIPS = 390.1
 

Dayman1225

Senior member
Aug 14, 2017
790
32
96
Time to revive this thread with the new Atom Roadmap!


A new core every two years, Tremont in 2019, Gracemont in 2021 and “Next Mont” (month is a typo) in 2023.

Tremont is on 10nm, wonder if Gracemont shrinks to 7nm or if they will wait for the Next Mont for 7nm.
 
Sep 9, 2017
58
9
41
Time to revive this thread with the new Atom Roadmap!


A new core every two years, Tremont in 2019, Gracemont in 2021 and “Next Mont” (month is a typo) in 2023.

Tremont is on 10nm, wonder if Gracemont shrinks to 7nm or if they will wait for the Next Mont for 7nm.
I think it's likely for Gracemont to be on 7nm, which makes more sense as TSMC is already working on 5nm and Intel needs to counter that.
 

you2

Diamond Member
Apr 2, 2002
4,363
11
91
How does the newest atom perf compare to newest snapdragon ?
 

Bouowmx

Senior member
Nov 13, 2016
821
9
116
Intel Pentium N5000/J5005 is between the Qualcomm Snapdragon 835 (Cortex-A73) and 845 (A75) in Geekbench 4 single-core (low-mid 2000s). Huawei Kirin 980 (A76) scores ~3360.
 
Apr 27, 2000
10,764
520
126
10nm Atom in 2019 huh. Hmmm, very interesting.

2021 for Gracemont on 7nm also seems likely. Gracemont may be used as a pipecleaner for the new EUV process. Might even be Jan 2021 if Intel is sufficiently aggressive. Anyone know when Tremont products will start hitting the market?
 

Dayman1225

Senior member
Aug 14, 2017
790
32
96
Time to revive this thread with the new Atom Roadmap!


A new core every two years, Tremont in 2019, Gracemont in 2021 and “Next Mont” (month is a typo) in 2023.

Tremont is on 10nm, wonder if Gracemont shrinks to 7nm or if they will wait for the Next Mont for 7nm.
Forgot to mention something quite interesting notes under Gracemont

“Vector Perf”, I wonder if this means Atom will finally see some sort of AVX instruction set?


10nm Atom in 2019 huh. Hmmm, very interesting.

2021 for Gracemont on 7nm also seems likely. Gracemont may be used as a pipecleaner for the new EUV process. Might even be Jan 2021 if Intel is sufficiently aggressive. Anyone know when Tremont products will start hitting the market?
I’d say late 2019. Lakefield being one of those products.
 
Apr 27, 2000
10,764
520
126
Forgot to mention something quite interesting notes under Gracemont

“Vector Perf”, I wonder if this means Atom will finally see some sort of AVX instruction set?
Intel has done it before as a part of Phi. Kinda makes you wonder where they'll use Gracemont. Also I missed it, but the 'NEXT' Month is a funny typo there. Heh heh. Who proof-reads those slides?

I’d say late 2019. Lakefield being one of those products.
Well well. It'll be interesting to see which makes it out first: Lakefield or IceLake U/Y.
 
Dec 17, 2008
1,845
6
91
Time to revive this thread with the new Atom Roadmap!

A new core every two years, Tremont in 2019, Gracemont in 2021 and “Next Mont” (month is a typo) in 2023.

Tremont is on 10nm, wonder if Gracemont shrinks to 7nm or if they will wait for the Next Mont for 7nm.
Wasn't this the original plan before 14nm cause no new advancement for you are in foundry progression hell? Remind me but wasn't the goal to have a new atom core ever 2 years roughly?
 

kenofstephen

Junior Member
Jan 18, 2008
4
0
66
Okay, here are the benchmarks! I've included my own Pentium J4205 (Apollo Lake) results where available and calculated the performance difference. Please note that the J4205 results are from Nov 2016 and used a different version of Windows as well as a much older graphics driver.

I observed the J5005 clock frequencies during an MT run of Cinebench R15. The CPU starts out at 2.7 GHz on all cores, but after ~1/3 of the test it settles down on alternating between 2.4 and 2.5 GHz.

It looks like the J5005 is 30-35% faster than the J4205 on average. I haven't done any extensive comparisons of operating frequency, but I'm guessing the J5005 runs at roughly 5% higher frequency. So average IPC improvement over Apollo Lake is probably in the 25-30% interval. Pretty impressive. Using Geekbench 4 for comparison, the J5005 seems remarkably close to the IPC of the Nehalem-based Core i5-760. Not spectacular performance for 2018, but I'd finally call the "Atom" pretty usable. This NUC, with a 256GB Crucial MX100, definitely doesn't feel slow. Single core performance appears to hover around 4 times that of the original Atom N270 (1.6 GHz) released in 2008.

After the table below, I've included URLs to some of the results (Geekbench/3DMark).



Result URLs

Geekbench 3:
J4205
J5005

Geekbench 4:
J4205
J5005 (new result with latest version of Windows)

3DMark Firestrike v1.1:
J4205
J5005


I spent an hour trying to get BOINC Manager to add the project, but it would always fail with a message along the lines of "Couldn't add the project. Please try again later". Yay... Tried two different BOINC versions and even attempted using boinccmd. Got a different error message there and decided to give up.
Hello! I admire your work a lot, may I know how I may send you a private message? Just some questions to ask about mini PC. Thanks!
 


ASK THE COMMUNITY

TRENDING THREADS