Intel Core i5-1035G1 and i5-1035G4: its iGPU would be half as powerful as Vega 11.
Contents
The Intel Comet Lake seem to be just around the corner, and it is that, as it happens with AMD, as we move forward in time, the leaks continue and their presentation date approaches. Today we have several interesting data that have been leaked, where we want to highlight the power of its iGPU, since it seems to be half as powerful as the current Vega 11.
The G series leaves shadows and several doubts
So far, the only thing we know about the 10th generation of Intel processors are various models of the G and U series, where we know many of its features. But many others remain to be revealed that are just as interesting for the end user, for this reason and based on today’s leaks, we will put white on black in various aspects, both of its performance and its architecture.

Several data have been leaked, specifically about the i5-1035G1 and i5-1035G4 CPUs, about which we already have a lot of relevant information. In the case of the i5-1035G1 and to refresh our memory, we are talking about a 4-core and 8-thread processor with a base frequency of 1 GHz and a Boost according to its SKUs for the 3.6/3.6/3.3 GHz cores.
In the case of the i5-1035G4, we are faced with the same number of cores/threads, but where the frequencies go up to 1.1 GHz and 3.7/3.6/3.3 GHz for its Boost. Having said this, we are faced with the position of questioning the data that we are going to show, since knowing the speeds that they will reach, the performance data is well below what it will finally be. In the same way we are going to reveal and comment on them.

The Intel Core i5-1035G1 has been seen at a frequency of 1 GHz on base, but only 1.19 GHz on Boost, giving SiSoftware scores of 59.98 Mpix/s on Processor Multi-media and 1.45 GB/s. s in Crytography Processor High Security, very low scores, where without a doubt the frequency of what we understand must be an ES is weighing them down.
We found a similar case in the data of the i5-1035G4, since in this case the score in Processor Multimedia was only 76.82 Mpix/s, where in this case the frequency was limited to 1.5 GHz, so there would still be 2.2 GHz left to explode in performance. In addition, we can anticipate the first data of the cache hierarchy, since they will carry in both cases 512 KB of L2 and 6 MB of L3, a considerable leap forward that shows the paradigm shift that we will find not only in Comet Lake but especially in Ice Lake.

iGPU data is much more interesting
The CPU performance data is not very interesting, given the low speed at which the tests have been passed, so it is impossible to compare it to that of other CPUs, but the iGPU data is more extrapolated and we will explain why. what do we discuss this
The data pertains to the i5-1035G1’s iGPU, which appears to have been limited to 1 GHz, thus 40C and 320 SP, giving a ratio of 8 SP per UC. They will share 1 MB of dedicated L2 and apparently a consumption of 15 watts (only GPU obviously). SiSoftware has once again been the one to reveal the performance data, where under GP (GPU) Scientific Analysis at normal precision this Intel UHD iGPU has achieved 70.61 GFLOPS.

If we compare it with Vega 11, its current rival, it has achieved an average of 147.36 GFLOPS in the same test. This means that AMD’s iGPU is 108.69% faster than Intel’s. Catastrophic result we can think, but there is a handicap that can change this perception: Vega 11 has 704 SP, while the new iGPU Gen 11 (in theory, since it could be Gen 9.5 too) from Intel in this CPU model only has 320 SP, that is, the iGPU of Lisa Su’s has 120% more SP than that of Swan’s, where on top of that they also have a higher frequency.
Will Intel finally include its Gen 11 or will it continue with Gen 9.5 at Comet Lake?
This makes us think that Intel would be above, or at least equal to, AMD in iGPU if it managed to match the number of SPs, something that seems quite likely if this 10th generation finally carries the new Gen 11 iGPUs. To situate ourselves better, the iGPU Intel’s fastest currently is the Iris Plus 655, which packs 48 UCs, 384 Shaders, and clocks between 1050 and 1200 MHz.
The question then would be simple: will Intel be able to include in the same 14 nm +++ manufacturing process an iGPU with a higher number of UCs/Eus under the Xe Gen 11 architecture? or maybe the data offered is the maximum it can offer under GT3e 9.5? In a short time we will know
