Section by Andrei Frumusanu

SPEC2017 and SPEC2006 Results (15W)

SPEC2017 and SPEC2006 is a series of standardized tests used to probe the overall performance between different systems, different architectures, different microarchitectures, and setups. The code has to be compiled, and then the results can be submitted to an online database for comparsion. It covers a range of integer and floating point workloads, and can be very optimized for each CPU, so it is important to check how the benchmarks are being compiled and run.

We run the tests in a harness built through Windows Subsystem for Linux, developed by our own Andrei Frumusanu. WSL has some odd quirks, with one test not running due to a WSL fixed stack size, but for like-for-like testing is good enough. SPEC2006 is deprecated in favor of 2017, but remains an interesting comparison point in our data. Because our scores aren’t official submissions, as per SPEC guidelines we have to declare them as internal estimates from our part.

For compilers, we use LLVM both for C/C++ and Fortan tests, and for Fortran we’re using the Flang compiler. The rationale of using LLVM over GCC is better cross-platform comparisons to platforms that have only have LLVM support and future articles where we’ll investigate this aspect more. We’re not considering closed-sourced compilers such as MSVC or ICC.

clang version 8.0.0-svn350067-1~exp1+0~20181226174230.701~1.gbp6019f2 (trunk)
clang version 7.0.1 (ssh://git@github.com/flang-compiler/flang-driver.git
 24bd54da5c41af04838bbe7b68f830840d47fc03)

-Ofast -fomit-frame-pointer
-march=x86-64
-mtune=core-avx2
-mfma -mavx -mavx2

Our compiler flags are straightforward, with basic –Ofast and relevant ISA switches to allow for AVX2 instructions. Despite ICL supporting AVX-512, we have not currently implemented it, as it requires a much greater level of finesse with instruction packing. The best AVX-512 software uses hand-crafted intrinsics to provide the instructions, as per our 3PDM AVX-512 test later in the review.

For these comparisons, we will be picking out CPUs from across our dataset to provide context. Some of these might be higher power processors, it should be noted.

SPECint2006

SPECint2006 Speed Estimated Scores

Amongst SPECint2006, the one benchmark that really stands out beyond all the rest is the 473.astar. Here the new Sunny Cove core is showcasing some exceptional IPC gains, nearly doubling the performance over the 8550U even though it’s clocked 100MHz lower. The benchmark is extremely branch misprediction sensitive, and the only conclusion we can get to rationalise this increase is that the new branch predictors on Sunny Cove are doing an outstanding job and represent a massive improvement over Skylake.

456.hmmer and 464.h264ref are very execution bound and have the highest actual instructions per clock metrics in this suite. Here it’s very possible that Sunny Cove’s vastly increased out-of-order window is able to extract a lot more ILP out of the program and thus gain significant increases in IPC. It’s impressive that the 3.9GHz core here manages to match and outpace the 9900K’s 5GHz Skylake core.

Other benchmarks here which are limited by other µarch characteristics have various increases depending on the workload. Sunny Cove doubled L2 cache should certainly help with workloads like 403.gcc and others. However because we’re also memory latency limited on this platform the increases aren’t quite as large as we’d expect from a desktop variant of ICL.

SPECfp2006(C/C++) Speed Estimated Scores

In SPECfp2006, Sunny Cove’s wider out-of-order window can again be seen in tests such as 453.povray as the core is posting some impressive gains over the 8550U at similar clocks. 470.lbm is also instruction window as well as data store heavy – the core’s doubled store bandwidth here certainly helps it.

SPEC2006 Speed Estimated Total

Overall in SPEC2006, the new i7-1065G7 beats a similarly clocked i7-8550U by a hefty 29% in the int suite and 34% in the fp suite. Of course this performance gap will be a lot smaller against 9th gen mobile H-parts at higher clocks, but these are also higher TDP products.

The 1065G7 comes quite close to the fastest desktop parts, however it’s likely it’ll need a desktop memory subsystem in order to catch up in total peak absolute performance.

SPEC2006 Speed Estimated Performance Per GHz

Performance per clock increases on the new Sunny Cove architecture are outstandingly good. IPC increases against the mobile Skylake are 33 and 38% in the integer and fp suites, though we also have to keep in d mind these figures go beyond just the Sunny Cove architecture and also include improvements through the new LPDDR4X memory controllers.

Against a 9900K, although apples and oranges, we’re seeing 13% and 14% IPC increases. These figures likely would be higher on an eventual desktop Sunny Cove part.

SPEC2017

SPECint2017 Rate-1 Estimated Scores

SPECfp2017 Rate-1 Estimated Scores

SPEC2017 Rate-1 Estimated Total

The SPEC2017 results look similar to the 2006 ones. Against the 8550U, we’re seeing grand performance uplifts, just shy of the best desktop processors.

SPEC2017 Speed Estimated Performance Per GHz

Here the IPC increase also look extremely solid. In the SPECin2017 suite the Ice Lake part achieves a 14% increase over the 9900K, however we also see a very impressive 21% increase in the fp suite.

Overall in the 2017 suite, we’re seeing a 19% increase in IPC over the 9900K, which roughly matches Intel’s advertised metric of 18% IPC increase.

Security Updates, Improved Instruction Performance and AVX-512 Updates Power Results (15W and 25W)
POST A COMMENT

261 Comments

View All Comments

  • tamalero - Monday, August 5, 2019 - link

    the 3d mark VRS test graph is very confusing. Reply
  • MASSAMKULABOX - Tuesday, August 6, 2019 - link

    These chips are priced quite high esp the i3's the Dualcore is outrageous (20$ discount for half the cores and threads?).They will only sell in limited qtys at such high prices . Bring on the 14nm equiv at sane prices.
    what discrete gfx does the IGPU equate to? gt1030 vs 2200g vs 3400ge??
    Reply
  • Zhentar - Tuesday, August 6, 2019 - link

    The bit per cycle throughput of REP STOS is really only half of the picture - the startup latency matters a lot too! On my Skylake, I've measured a minimum latency of 29 cycles* (any REP STOS from 1 byte to 128 bytes takes 29 cycle, then it starts going up from there). Some compilers make heavy use of it even for small stores/copies (VC6, first and foremost, but also the .NET JIT for stack zeroing), so it can be pretty important to performance in some scenarios.

    *I am no Agner Fog... nor anywhere close to him, this measurement should be taken with a grain of salt ;-)
    Reply
  • ballsystemlord - Wednesday, August 7, 2019 - link


    @Ian
    Sunny Cove SIMD chart and others have many asterisks, why?

    "POPCNT Microcode 50% faster than SW (under L1-D size)"
    What does this mean? The CPU uses microcode and HW, not SW, AFAIK, much less use it to implement the POPCNT instruction.
    Reply
  • ballsystemlord - Thursday, August 8, 2019 - link

    Spelling and grammar errors:
    Insert: I've been commenting corrections for at least 1 year now, if you guys want me to change the format, or have any request regarding phrasing, etc., just ask. Thanks for your work!

    "L3 latencies look similar, however we'll dwell into that just in a bit."
    I think you ment delve:
    "L3 latencies look similar, however we'll delve into that just in a bit."

    "IPC increases against the mobile Skylake are 33 and 38% in the integer and fp suites, though we also have to keep in d mind these figures go beyond just the Sunny Cove architecture and also include improvements through the new LPDDR4X memory controllers."
    Missing percent sign and stray "d".
    "IPC increases against the mobile Skylake are 33% and 38% in the integer and fp suites, though we also have to keep in mind these figures go beyond just the Sunny Cove architecture and also include improvements through the new LPDDR4X memory controllers."

    "...which did not go down to well."
    You mean "too" not "to":
    "...which did not go down too well."

    "...we actually see a number of key microarchitectural improvements bubble up through in our SPEC testing."
    Excess "in":
    "...we actually see a number of key microarchitectural improvements bubble up through our SPEC testing."
    Reply
  • alysdexia - Thursday, November 28, 2019 - link

    I think you mean Baalsystemlord or Bàqhàlsýstemlord.
    Grammar is a barbarism of ghrammatics, another word for composition or handwriting, font, format, spacing, the looks, whereas you meant diction/lecsis or register, vocabulary, declension, placement, spelling, the meaning.
    you guys -> ye
    Reply
  • nils_ - Monday, August 19, 2019 - link

    I noticed on Intel Ark that Ice Lake CPUs do not have TSX instructions enabled / available. THis is interesting, since I believe TSX has also been used for some of the spectre attacks. Reply
  • ikjadoon - Thursday, September 12, 2019 - link

    On 'Gaming Results (15W and 25W)', the graph is colored incorrectly. "On" and "Off" have the same color...and the legend colors are somehow not related to the graphic? Reply

Log in

Don't have an account? Sign up now