SSE4 on the Go

Obviously one of the strengths of the Penryn architecture is its SSE4 support. We ran Intel’s TMPEGEnc benchmark, which does take advantage of SSE4 during a movie transcode, and the performance gap widened:


TMPEG

The mobile Penryn based system, because of its SSE4 support, can complete the encoding test in 31% less time than the Merom system. The impact of SSE4 is even more pronounced on the mobile side because of the fact that the mobile chips are already so starved for bandwidth.


VirtualDub

The performance gap widened when running a SSE4 optimized VirtualDub 1.7.2 DivX encode, here Penryn would offer a greater than 40% increase in performance.

These SSE4-optimized situations are far more rare than the 1 - 8% increases we saw elsewhere, the point being that should application support develop, Penryn could do much better. Honestly though, we don’t expect a critical mass of SSE4 applications anytime soon, these sorts of things take a long time to materialize.

Mobile Penryn vs. Merom: Performance Final Words
POST A COMMENT

18 Comments

View All Comments

  • metacircular - Friday, January 11, 2008 - link

    Thanks for the review, but could the reviewers comment on the temperature differences between the two processors? I'm assuming lower voltage will result in less heat, but it would be nice to see some numbers. Reply
  • legoman666 - Thursday, January 10, 2008 - link

    It's a shame about that DAT not really being useful. It seems like their ideas were hampered by the fact that applications always bounce from 1 core to another. Looks/sounds like a good idea though. Reply
  • Jussi - Wednesday, January 09, 2008 - link

    Thanks for an interesting article, Penryn it seems like a nice, although not essential refresh.

    I'd like to nitpick the SSE4 performance numbers. You state that in case of Virtualdub "Penryn would offer a greater than 40% increase in performance". I find this to be incorrect.

    Let performance or speed (v) be defined as v = work unit / time. v1 is the speed of Merom and v2 is the speed of Penryn. Comparing Penryn to Merom = v2/v1 => t1/t2. Using your numbers 47.3s / 28.2s = 1.677 => 68% better performance.

    It would be correct to state that using Penryn takes about 40% less time to do the job, but that is not what the article states.
    Reply
  • puffpio - Tuesday, January 08, 2008 - link

    Could I just swap chips out? Reply
  • channelv - Wednesday, January 09, 2008 - link

    Yup, the D630 uses the 965 chipset (Santa Rosa). But you'd better wait for a BIOS update from Dell first before you think about putting a Penryn in there - I'd expect those to roll out almost any day now, but by Feb. for sure. Reply
  • Mr Alpha - Tuesday, January 08, 2008 - link

    Have I understood it correctly, that you are comparing the SSE4 version of DivX on Penryn with the same SSE4 version of DivX on Merom? Shouldn't you be comparing the SSE4 version of DivX on Penryn to the SSE3 version of DivX on Merom? Reply
  • mi1400 - Tuesday, January 08, 2008 - link

    Electric surges drain battery faster. When a device bulb, motor etc starts a peak is occured in load and then load comes to a lesser and steady value. Intel Dynamic Acceleration may result in same for battery. Reply
  • Mgz - Monday, January 07, 2008 - link

    how about the Deep Power Down (C6) stage that penryn brings ? Any test on this feature? Reply
  • coolme - Monday, January 07, 2008 - link

    Good review, but just to note.

    Merom (part of santa rosa) supports Intel Dynamic Acceleration, (although T7xxx series only) and since you guys weren't aware of that, the cinebench benchmark (and other benchmarks) with the T7800 might have a skewed score.
    Reply
  • duploxxx - Monday, January 07, 2008 - link

    Nice to see that the t9xxx series will step up the performance again and battery life. Altough we can already say that Intel dominated this part of the mobile market for a long time with t7xxx it is now sure that it will increase that even more.

    what i really want to no is what about low end and midstream, how good are these t2xxx - t5xxx - t7xxx series compared to each other and what happens when you put a rather cheap turion class next to those t2 and t5 series.

    @anand i think you would hit a readers max out of such a review, but then again it would take a lot of time and effort to get a nice compare with equal hardware parts and price....

    Reply

Log in

Don't have an account? Sign up now