Personal View site logo
AMD Ryzen aka Zen CPU - Broadwell-E Performance at a Lower Price Point
  • 76 Replies sorted by
  • Ugh, the benchmarks should be sorted by results rather than arbitrarily. It's easy to look at those results and think it came in second in all of the tests when it really came in second, sixth, and third.

  • @eatstoomuchjam The multi-threaded price/performance ratio is in first place by a huge margin. This thing is showing the same performance as a CPU that costs 3-4 times as much. Seems very promising for a workstation, IMO. Formal benchmarks will be arriving in about a week.

    Price-Performance

  • @tron I was only referring to the confusing order of the results in the graphs. The statement I made was factually correct. I made no reference to price.

  • I will almost certainly (once independent benchmarks confirm the current "rumored" ones) be one of the early adopters buying an R7 1800X. I would have upgraded my PCs a long time ago had there been actual progress in performance per price, I am happy to see that finally such seems to materialize. Upgrading from 4 to 8 cores will definitely give me a noticeable speedup.

  • I pre-ordered the 1700. It will probably be a couple weeks before I have my hands on the MB to finish the build. I'll be performing some benchmarks on render times vs. an i7-4790 system once it's up and running.

    @karl Please post your results here if you do any comparisons on render times. Thx.

  • Benchmarks have begun showing up.

    http://www.anandtech.com/show/11170/the-amd-zen-and-ryzen-7-review-a-deep-dive-on-1800x-1700x-and-1700

    The CPU has a traditional uArch and does well, especially compared to last generation AMD, and a new high-perf core will be a feather in their cap. We see a lot of benchmark results where AMD is clearly equal or above Intel's HEDT parts in both ST and MT. However there are a few edge cases where AMD is lacking behind 10-20% still, even to Broadwell. These edge cases are difficult to anticipate, and can stem from unoptimized code. One of the benefits of Intel's big R&D juggernaut is the ability to process those edge cases, through prefetch, memory algorithms, and extensive testing. So despite the best will, there's still a large element to having a substantial budget to hire 300+ more engineers to cater for that, which is something AMD wasn't able for Zen or Ryzen.

    But Senior Engineer Mike Clark says he knows where the easy gains are for Zen 2, and they're already working through the list. A question is then if Intel continues at 5% performance gains clock for clock each year, can AMD make 5-15% and close the gap?

    Price for Performance is still one of AMD's strong points. For the multitude of tests where that $499 1800X is able to match or beat a $1049 i7-6900K, it directly translates to a 2x in price/performance. Intel still has the plus on the ecosystem and chipset, but the argument is around how many users actually do what that extra IO.

  • @eatstoomuchjam

    Media fully focused on premium models, yet mass market is i3 and i5 chips (for desktops). 6-8 cores that are so heavily discussed in media take 1-2% of consumer marker.

  • Applications benchmarks

    image

    image

    image

    So, absolutely no reason to update for i7 4 core CPUs owners.

    Add to this lack of QuickSync and for fast work you can get slower computer.

    http://www.ixbt.com/cpu/amd-r7-1800x.shtml

    sample845.jpg
    648 x 199 - 37K
    sample846.jpg
    646 x 198 - 38K
    sample847.jpg
    650 x 198 - 32K
  • Computer industry has been price gauging consumers/enthusiasts due to youtube/amazon monopoly-based Guerilla marketing personal recommendations/blogs... go back to Core 2 Duo days to see the trend lines

    with scores of willing shills who are rarely critical of products as they NEED YOU TO CLICK THE LINK to make a few pennies from amazon, etc, manufacturers LOVE bloggers/youtubers, who can be easily conned to HYPE their product launches to gain access in ways they never had in the past, especially compared to when they actually used to pay for magazine ads,

    After all, modern Smartphones come with "octocores" and 4K video cameras now... shouldn't a PC desktop/laptop CPU?

    regardless, I am happy AMD now has competitive 8 core product for Apple to launch new Mac Pro in new CPU and GPU options, to push Intel to make 8 cores mainstream for desktops...

    I seriously thought about cost for me to refresh one of my intel machines to AMD Ryzen and VEGA, and concluded something really SEXY...

    insufficient motherboard expansion port options compared to X99 and minimal performance bump in creative real-world compared to Z270 with future optane options, as my replacement targeted intel 6 core with GTX GPU and TOTAL LACK of Thunderbolt 3 motherboard options for Ryzen at this point which is sorely needed for modern audio interfaces for my upgrade to occur...

    Yep, something as simple as 10 sata ports and thunderbolt/usb c expansion matter to geeks like us...

  • with scores of willing shills who are rarely critical of products as they NEED YOU TO CLICK THE LINK to make a few pennies from amazon, etc, manufacturers LOVE bloggers/youtubers, who can be easily conned to HYPE their product launches to gain access in ways they never had in the past, especially compared to when they actually used to pay for magazine ads,

    Well, in capitalism you have only two other options - paid subscriptions (for most sites it is not option), and donations (normally you won't survive on them even for big sites).

  • Re: Well, in capitalism you have only two other options - paid subscriptions (for most sites it is not option), and donations (normally you won't survive on them even for big sites).


    Ah yes, it's late here... I did forgot to go home and watch and make reaction videos to next Disney Star Wars / Marvel comic movie or the next sponsored how to build a RAID 0 Seagate 1 Petabyte storage array with an 8K Red camera...

    as in all things in life and wife, buyer beware

  • More benchmarks

    image

    image

    image

    image

    image

    sample858.jpg
    681 x 248 - 32K
    sample859.jpg
    680 x 245 - 31K
    sample860.jpg
    678 x 241 - 32K
    sample861.jpg
    675 x 248 - 32K
    sample862.jpg
    676 x 247 - 32K
  • Quick update on Ryzen... performance is heavily locked to ram speed, so some of the early benches floating around aren't even close to what this thing can do due to some early bios revs not supporting high memory clocks.

    "This is a very important technical info that most of us probably had no idea.

    Traditionally, RAM speed is not linked in anyway to how fast the CPU communicates, it's a separate element. However, Ryzen's design is very much linked, in that the Northbridge of Ryzen runs at 1:2 clock speeds of the memory controller.

    The result is unlike other CPU designs, where faster RAM often yield no gaming advantages, with Ryzen, it directly boosts performance by boosting CPU communication clock speeds. Irrespective of the added RAM bandwidth having an influence on the app/game.

    As new bioses roll out and RAM speeds increase, Ryzen performance will grow."

    More info here: https://www.reddit.com/r/Amd/comments/5xjmau/ryzens_data_fabric_is_locked_12_ram_speed_faster/#bottom-comments

  • My 1700 is currently clocked at 3.6 Ghz (1.2v) on all cores with the stock wraith cooler and DDR4 @ 2933 Mhz. Temps are 30C at idle and 55C during long renders. Having run different Ram tests at a variety of speeds, I can vouch that the above statement is absolutely correct with regard to the Ram speed performance impact. I did a quick bench in CPU-Z and was surprised to see this thing actually clobbers the 10 core i7-6950X in multi-thread.

    Real World Test

    Magix Vegas Pro 14 - 30 Sec Render, 4K Project (basic color correction) w/GPU On

    i7-4790 @ 3.6-4.0Ghz (stock) - 3 Minutes 6 Seconds

    R7 1700 @ 3.6Ghz - 2 Minutes 3 Seconds

    So I'm seeing around a 50% decrease in render times vs. the 4c/8t Haswell. Preview speeds have also increased dramatically. Not too shabby for a $329 CPU, very happy thus far.

  • @Tron

    As far as I remember testers used highest speed officially supported memory.

    Also Sony Vegas can have issues working with QuickSync or your config is wrong.

  • @Vitaliy_Kiselev I triple-checked my settings on both machines. Same RX 480 gpu, same source project off an external USB drive, both with 16 GB Ram and Win 10 on SSD.

    Honestly, I think the QuickSync feature is becoming less relevant on workstations with the arrival of "Coremageddon." The brute force of 8c/16t provides better render times and higher image quality overall. That could be the reason Intel doesn't even offer QS on their 6+ core CPUs.

    For mobile systems I think QS will continue to be king, unless AMD starts dropping 8 core CPUs in that segment as well. The Ryzen chip sips very little power when down-clocked to 2.8 GHz. It's like 45W which is the same as the i7-6700HQ.

  • @Tron

    QuickSync can have issues due to video config (will be off if you disabled build in GPU).

    Also Vegas can be not the best program to test rendering speed as it is known for horrible encoding times.

  • OK, good info. I'll jump into bios on the Intel machine to check the iGPU state and rerun the test if it's disabled. My understanding is that GPU acceleration in Vegas is via Open CL, so not sure if it will make a difference either way.

    The only other resource I use for my comps is After Effects but that is version CS5, so optimization may be unrepresentative of the current 2017 CC version. I'll run a quick test to see how it fares.

  • Encoding does not work good via CUDA or OpenCL. It is always uses dedicated block like QuickSync or build in GPU for NVidia.

  • Example how capitalism works:

    All current Ryzen chips have 8 cores, some of 4 and 6 core chips have disabled cores due to defects, but most are disabled to make separate price niche.

  • Theyhaverisen.

    Im waiting for new opeton parts with 64 cores, and 128 threads. Thats the niche im more interested in. Well thats what the road map shows and where the big bucks are. For AMD and for end user.

    Also, i have notice as well that this processor is very memory speed dependent and with each L3 cache separated for each processor, this may end in random data access errors from different programs that are not optimized for this workflow. This is an issue that can be easily corrected in time with software fixes and updates from programs.

    Even though having not optimized software for this processors, they behave extremely well for a new architecture. This is solid rock good performance. I cant wait for the opteron parts to arrive, on 2 way socket. Should be an animal of processors, at a good speed.

    I also think its scalar future is quiet good as its older sever grand grand dad the opteron, at a reasonable price.

    amd-zen-naples-two-socket.jpg
    650 x 334 - 53K
  • AMD suggest to use DDR4-3200 or DDR4-3500 modules and only one pair of identical modules. Each motherboard can have issues with some modules at such frequencies and 4 modules result in speed drop.

  • Just a quick follow-up... I enabled the iGPU on the i7-4790 to do a quick comparison against the discrete GPU (Open CL) to see if QuickSync was supported in Vegas. Unfortunately it was drawing random black frames and choking something fierce with projected render times of 8 minutes+ on a 30 second project. Confirmed that the Intel drivers were the latest. So it appears the best performance available on that architecture w/Vegas is with a discrete GPU, which still ends up being around 50% slower than the Ryzen chip clocked at 3.6Ghz. I can't speak for Skylake and Kabylake experience but it looks like similar results are being seen against Broadwell-E CPUs as evidenced in this comparison vid at the 5:10 mark -

  • Unfortunately it was drawing random black frames and choking something fierce with projected render times of 8 minutes+ on a 30 second project.

    Before making conclusions it is better to check motherboard firmware and consult Magix support.