Adreno 530 faster than Intel HD530 ?

mohit9206

Golden Member
Jul 2, 2013
1,381
511
136
So i was browsing this site which lists the GFlops of various mobile, Intergreted and desktop GPUs and i was surprised when i saw that the mobile gpu Adreno 530 in Snapdragon 820/821 has more GFlops than HD530 graphics in Intel Skylake processors.
Intel HD530 having peak 441 gflops at 1150 mhz while Adreno 530 having peak 498 GFlops at 624mhz in SD820 and 519 gflops at 650 mhz in SD821.
Now i know that theoretical gflops don't really mean much but its still surprising that mobile gpu has more gflops than Intel desktop igpu.
http://kyokojap.myweb.hinet.net/gpu_gflops/
Maybe its time for Intel to really push the igpu forward as its kinda embarrassing to become outclassed by a smartphone gpu?
 

VirtualLarry

No Lifer
Aug 25, 2001
56,570
10,202
126
Don't forget, mobile GPUs usually operate in FP16, while PC GPUs uses FP32.

I know that this has been brought up multiple times, when comparing mobile iGPUs with desktop iGPUs.
 

mohit9206

Golden Member
Jul 2, 2013
1,381
511
136
These new mobile GPUs from last couple of years have already overtaken Xbox 360 and ps3 gpu in gflops so it wouldn't be surprising if they actually are faster than Intel igpu even in the same fp category.
 

LTC8K6

Lifer
Mar 10, 2004
28,520
1,575
126
Also, Adreno 530 is the top end, while HD530 is midrange.

In fact, there are earlier Gen 7.5 and Gen8 HD graphics without edram that have numbers much higher than Adreno 530.
 

DidelisDiskas

Senior member
Dec 27, 2015
233
21
81
It is interesting that the hd530 at that page is listed as having 345.6~441.6 gflops, while the iris 6100 is listed at 806.4~844.8. I've tried them both and TF2 ran at about the same performance. What causes this kind of discrepancy between the theoretical and the real world performance?
 
Last edited:

Face2Face

Diamond Member
Jun 6, 2001
4,100
215
106
It is interesting that the hd530 at that page is listed as having 345.6~441.6 gflops, while the iris 6100 is listed at 806.4~844.8. I've tried them both and TF2 ran about at about the same performance. What causes this kind of discrepancy between the theoretical and the real world performance?

I've noticed the source engine runs very much the same on a lot of Intel iGPUs, especially at 720p. I've tested this myself with my N3700 vs. m3-6Y30, and the difference is very slight at 720p, but at 1080p the Core m3-6Y30 sees a decent bump in performance over the N3700. I've done some benchmarks, comparing three models of Intel Compute Sticks - Bay Trail vs. Cherry Trail Vs. Core M.

Here's Dirt 3 and RE 5.

You'll see the Core m3-6Y30 with the HD 515 is fairly capable, and uses almost half the power of a X5-Z8330 (Cherry Trail)

https://www.youtube.com/watch?v=V2mZMAnm2YQ
https://www.youtube.com/watch?v=Uzt-akeOw5I