Samsung Exynos 5250 - production starting in Q2 2012

  • Thread starter Deleted member 13524
  • Start date
One thing people seem to miss from the datasheet which is new:

1.4.10.2 Cortex-A5
 Low power co-processor unit
 ARM Cortex-A5 core processor uses ARMv7-A architecture
 16 KB instruction cache and 16 KB data cache
Click to expand...
But I doubt they can make full use of this. The complexity in scheduler and everything else is just too much for at the moment. The Linux kernel isn't at that point yet right now.


Edit: Oh god the user manual is beautiful. I wish they had released this for the 4412, it would have saved me a damn lot of time in kernel hacking...
 
Last edited by a moderator:
Jubei said:
I thought A6X was 2X A5, not A6?
Click to expand...

I misspoke, but you are also wrong...the A6X is x2 the A5X. I'm guessing its only about 25-30% quicker than the A6, so the 5250 CPU might still easily outperform it.

Samsung claims 1,7 GHz Exynos 5 has 2X better performance than a dual 1,4 GHz Cortex A9
Click to expand...
Sounds likes a similar claim to that apple has made for the difference between its 800mhz A5 in the iphone4s V's the 1.3ghz A6 in the iphone5
 
My mistake i forgot an X there somewhere :)

In the same PDF Samsung is claiming Mali-T604 to have 2X better performance than Exynos 4 GPU (im assuming 4412) wich should make it faster than A5X but not quite up there with A6X?
 
Jubei said:
My mistake i forgot an X there somewhere :)

In the same PDF Samsung is claiming Mali-T604 to have 2X better performance than Exynos 4 GPU (im assuming 4412) wich should make it faster than A5X but not quite up there with A6X?
Click to expand...

I'm surprised if it is only getting x2. Arm stated it would be x5 of mali400, and the exynos5 is clocking the gpu 25% quicker than exynos4.
A5x is getting around x1.5 of the Exynos4. assuming both "X2" claims are correct, then A6X should still have a x1.5 advantage.

If true, arm should surely be disappointed that the first outing of their next gen is still easily beaten by a gpu core technology that has been in production for well over 2 years, and in all likelihood, one that is clocking slower
 
For the 2.1 onscreen test: as Nebuchadnezzar said it's in 2500*1600 which is quite a bit higher than 2048*1536 for iPad3 or iPad4. Not much of an indication anyway since onscreen means vsynced at 60Hz.
 
tangey said:
I'm surprised if it is only getting x2. Arm stated it would be x5 of mali400, and the exynos5 is clocking the gpu 25% quicker than exynos4.
A5x is getting around x1.5 of the Exynos4. assuming both "X2" claims are correct, then A6X should still have a x1.5 advantage.
Click to expand...

Depends how they measured. Both the T604 as the Mali400MP4 in the last revision of the 32nm Exynos4 seem to clock at comparable frequencies. If the comparison ignores native resolution an it's 1280 vs. 2560, the claim is anything but absurd for a marketing whitepaper.

If true, arm should surely be disappointed that the first outing of their next gen is still easily beaten by a gpu core technology that has been in production for well over 2 years, and in all likelihood, one that is clocking slower
Click to expand...

Some guesswork: assume T604 has 4*SIMD16 (4 clusters), they might be counting some additional single FLOPs from SFUs to get those 72 GFLOPs peak at 533MHz. 1 TMU/cluster gives
2.132 GTexels/s peak theoretical fillrate.

Assuming the A6X contains a SGX543MP4@500MHz:

4 cores * 4 ALUs * 9 FLOPs * 0.5GHz = 72 GFLOPs
8 TMUs * 500MHz = 4.0 GTexels/s
64z * 500MHz = 32.0 GPixels/s

I don't dare to predict yet how the T604 will fare in GLBenchmark2.5, but if their driver/compiler should be good enough (which I don't see why it shouldn't) the fact that they might have "scalar" SIMDs might give them an efficiency lead over Series5XT. Adreno320 is also quite a bit slower in 2.1 against the SGX543MP4@250MHz, but it's a totally different story in 2.5.
 
tangey said:
I'm surprised if it is only getting x2. Arm stated it would be x5 of mali400, and the exynos5 is clocking the gpu 25% quicker than exynos4.
A5x is getting around x1.5 of the Exynos4. assuming both "X2" claims are correct, then A6X should still have a x1.5 advantage.

If true, arm should surely be disappointed that the first outing of their next gen is still easily beaten by a gpu core technology that has been in production for well over 2 years, and in all likelihood, one that is clocking slower
Click to expand...
The x5 claim came way before the 32nm 4412. The 400MP4 on the Note 2 is already running at 533MHz, that's basically double what the 4210 and ST-Ericsson versions run at. 5x Mali 400 at 266MHz is not absurd at all and pretty much in range of the possible.
 
Ailuros is correct. Now that i read through the pdf properly, Samsung was meauring it as 2x better than 4412 at wxga resolution. Im assuming its 4412 given that they mentioned cpu frequency of 1,4 GHz
 
Jubei said:
Ailuros is correct. Now that i read through the pdf properly, Samsung was meauring it as 2x better than 4412 at wxga resolution. Im assuming its 4412 given that they mentioned cpu frequency of 1,4 GHz
Click to expand...

Well to put it on another level I don't believe that a T604 would reach almost 10k frames in GLBenchmark2.5 (1080p offscreen) before I see it.
 
GLBenchmark from a leaked device: http://briefmobile.com/nexus-10-receives-the-glbenchmark-and-sunspider-treatment

Resized picture for the sake of sanity:


IRqJo.png


Pretty equal to the A6, but gets beaten in low-level benches. Also tied up with the A5X. I guess the it won't be a match for the A6X. On the other hand beats the Adreno 320 handily.

I wonder how drivers will improve it, it's still a new architecture. Remember that the Mali 400 got roughly 40% boost from early release to several months later through drivers.

It does keep its 2x performance over a ~440-533MHz Mali400. 5x 266MHz Mali400 seems to be roughly validated.
 
Last edited by a moderator:
I guess it's all semantics, but when I read in the original PR for T-604:-
"The Mali-T604 delivers up to 5x performance improvement over previous Mali graphics processors...."

My reasonable expectation is that this is on a clock to clock comparison, i.e. an archeiture improvement
 
Nebuchadnezzar said:
Pretty equal to the A6, but gets beaten in low-level benches. Also tied up with the A5X. I guess the it won't be a match for the A6X. On the other hand beats the Adreno 320 handily.
Click to expand...

Where does it beat adreno 320 handily? Even at briefmobile they said that adreno 320 was faster at GLbenchmark, not always but it wasn't beaten as you say it was. And when you compare scores from the fastest adreno 320 at glbenchmark it wins with nexus 10 at almost all offscreen tests.
 
I see that BriefMobile also ran SunSpider, that score is much worse than I expected. The Chromebook with the same chip is twice as fast. I can accept that Chrome on Android is not the same version as Chrome on ChromeOS but that is a huge difference for software from the same company.
 
Nebuchadnezzar said:
GLBenchmark from a leaked device: http://briefmobile.com/nexus-10-receives-the-glbenchmark-and-sunspider-treatment

Resized picture for the sake of sanity:


Pretty equal to the A6, but gets beaten in low-level benches. Also tied up with the A5X. I guess the it won't be a match for the A6X. On the other hand beats the Adreno 320 handily.

I wonder how drivers will improve it, it's still a new architecture. Remember that the Mali 400 got roughly 40% boost from early release to several months later through drivers.

It does keep its 2x performance over a ~440-533MHz Mali400. 5x 266MHz Mali400 seems to be roughly validated.
Click to expand...

Mali400MP4@533MHz = 1977 frames
Mail400MP4@440MHz = 1827 frames (latest Galaxy SIII score)
Mali400MP4@266MHz = 1475 frames (latest Galaxy SII score)

3313 / 1977 = 1.68x
3313 / 1827 = 1.81x
3313 / 1475 = 2.24x

With an up to 40% driver improvement in the future they'll slightly exceed the 2x times mark over the 533MHz Mali400MP4 under the presupposition that the latter won't see a single inch of addtional performance against the T604.

Thost 2x or 5x times metrics don't have to be for GLBenchmark-anything by the way. I'm sure it won't be too hard to find a case with high enough geometry load where a T604 could literally slaughter a 400MP4@533MHz.
 
You must log in or register to reply here.

Similar threads

F
Exynos 5450
2 3
Replies
43
Views
14K
Lazy8s
L
D
Creative ZMS-20 and ZMS-40
2 3 4
Replies
75
Views
22K
darkblu
D
M
Samsung Orion SoC - dual-core A9 + "5 times the 3D graphics performance"
16 17 18
Replies
358
Views
108K
Nebuchadnezzar
N
A
Holy Wall of Text: My impressions post Playstation 4 (PS4) unveil
2 3 4
Replies
70
Views
19K
(((interference)))
I
Back
Top