Intel Core i7-3720QM Review -- Mobile Ivy Bridge To The Test

Which Processor exactly are you testing??

The title says i7 3720qm, but your laptop specs say i7 3610qm, and so does your laptop performance page
 
Pretty impressive.
Although I would have liked to have seen other mobile CPU's on here, maybe an i5-2430M and a i7-2720QM...
 
SpecViewPerf is a set of engineering graphics benchmarks, so the SpecViewPerf results compare the performance of the the GPUs: the notebook NVidia 650M and the desktop NVidia 580. The commentary comparing the CPUs on this benchmark should be comparing the GPUs. The CPU should be relatively idle during this benchmark. (I think in each benchmark the CPU has a large model in memory and replays a stream of OpenGL API calls [to render the model] that are sent to the GPU, in the same order as it was recorded in a traced run of the corresponding software.) The 650M looks comparable to the desktop 580 for single-precision graphics work, even if it is hampered by reduced double-precision floating point units.

(Typo in first test system spec and later laptop spec: it reads 3620QM but Intel's site lists no such part, should it be 3720QM or 3610QM ? As far as I can see from the ASUS site, N56VZ is only available with the 3610QM. Did you get a new version, or put in your own chip?)

(Titling the two bargraphs of SpecViewPerf subtests with the same title and subtitle was confusing. Only at the bottom is there a small caption saying one was SolidWorks and the other was Maya.)
 
How do you explain the large difference in performance between the various desk top CPU’s which are all using the same GPU? How do we compare GPU’s when 14 of the desktop processors were all using the same GPU. Meanwhile what you are suggesting is that the GeForce 650M is faster than the GeForce GTX 580 in these “GPU” tests when armed with the i7-3720QM mobile processor.

Yes that was an obvious typo in the system specs, correct we put a 6 instead of a 7, fixed now. The title of the review and the graphs all state that the i7-3720QM was used so I cannot see how a simple typo create that much confusion.

Also again sorry that the Specviewperft graphs caused confusion.
 
Agreed, a GPU comparison would not be sufficient to explain all the differences. The top half of the order in the specviewperf solidworks test does not appear to be correlated with base or turbo freq, so it does not appear to test CPU compute performance. This top half order appears somewhat random within several categories for configurations with the same GPU. The socket 2011 processors with 4 memory channels are at ~12fps and the next group with 2 memory channels is all at ~10.5fps. Maybe they are constrained by the memory and I/O bandwidth required to copy data from memory to the GPU. If the laptop copies this data using PCIe3 between the CPU and GPU, that might explain how the laptop with PCIe3 but lower memory bandwidth, lower CPU frequency, and weaker GPU is able to outperform several desktop processors with PCIe2 but higher CPU frequency, higher memory bandwidth, and stronger GPU.
 
Back