Slashdot Mirror
AMD Unveils Ryzen Threadripper 1950X 16-Core and 1920X 12-Core Specs and Pricing (hothardware.com)
MojoKid writes: AMD first teased its Ryzen Threadripper series of high-end desktop (HEDT) processors back in mid-May, but is now sharing additional details on the first two products in the family. Both processors are based on the 14nm Zen core, make use of AMD's new Socket TR4 interface, support quad-channel DDR memory, and feature a total of 64 PCIe lanes. In addition, both processors will come from the factory unlocked. Ryzen Threadripper 1920X will have 12 Cores, 24 Threads, and 3.5/4.0 GHz (Base Clock/Precision Boost) clock speeds. Ryzen Threadripper 1950X will have 16 Cores, 32 Threads, and 3.4/4.0 GHz (Base Clock/Precision Boost) clock speeds. Pricing is set at $799 and $999, respectively, with availability in early August, though Dell's Alienware gaming PC division will have systems shipping with the new chip starting this month. AMD also put the new chips up against Intel's Core i7-7900X 10-core CPU in a Cinebench benchmark run in a video demo, and the 12-core Threadripper chip beats Intel's currently available Skylake-X chip handily, while the 16-core Threadripper outpaces it even further.
These sell at a steep discount to Intel's core i9, and are anywhere from $400 to $1000 cheaper. It's a no-brainer.
The plethora of undocumented opcodes present in these chips make them perfect for running Linux compilers. Some operations can run in 1/10 the time if properly optimised.
Ryzen IPC is better than even Broadwell-E, which is newer than your Haswell. I probably wouldn't upgrade a Haswell i7 for gaming either, at least not until the publishers start taking advantage of the higher core counts.
I know what I want for Christmas.
Such a shame my wife doesn't read /.
I've always been a fan of AMD's processors and still have used them in my Linux machine, but I had to go Intel for my gaming system for the last few upgrades. Granted I never spend this much on a CPU so I'll need to tech to "trickle down" to their budget line, but seeing a good performance option from AMD will be good again.
"People who think they know everything are very annoying to those of us who do."-Mark Twain
In a dwindling X86/AMD64 PC market, laptops* is where the volume is... Yet AMD has nothing for the Laptop Market in the Zen Class Architecture.
And in servers, while there may not be as much volume, is where the cream of the profits are.
While Zen Server parts (Epyc) look good on paper, it reamis to be seen if there will be Adoption from server makers, and demand from server purcharsers...
So, no laptop parts, to early for servers, coupled with so so results for enthusiasts desktop PC (great bang for buck, but performance is more or less even depending on workload) and crap processors for enterprise desktop (corporate parts without IGP? Really? I mean, REALLY?!?!?), is to early to be happy for AMD.
I hope they do well, I really do, for this will be good for all of us (even those of us using Apple gear, therefore, tied to Intel)...
But one thing is to hope, and quite another thing is reality, and is to early to know what reality looks like.
Just my two cents.
*** Suerte a todos y Feliz dia!
And with all these advances Windows is still single threaded when it comes to most of the functionality.
Get blown away in games. Sit for hours updating your OS.
How about you use a source that actually knows what they're talking about, and knows how to do reliable, repeatable testing?
Ryzen 5 1600X vs Core i5 6600K
Cost difference is negligible - a minor discount on either side will swing it. Nominal TDP is only 4W apart. And it's the high-end "ordinary consumer" part - the default recommendation for PC gaming. This is as close to an even comparison as you can get.
Across multiple graphics cards, across multiple games at different resolutions, AMD is competitive. Major wins on some games (Civilization 6), major losses on others (Rocket League), plenty of dead ties (GTA V), and a general trend of AMD doing better as resolution increases. No real oddities with uneven framerates - the 99th percentile framerate tracks the mean. AMD gets a small but consistent lead on synthetic benchmarks, and naturally scores overwhelming wins on multithreaded rendering.
Intel is fucked on all levels server / desktop and workstation / high end gaming. By cheaper AMD systems with more pci-e lanes at all levels.
Intel kaby-lake x is an said over priced joke that costs more and does less then the same chips on the desktop.
My citations are here [youtube.com] and here where the 2600k beats the newer Ryzen even further [youtube.com]?
Er what? In the 930 video, at 4:45 and 10:00: The lowest AMD was the stock Rzyen CPU R5 1500X which beat out the stock 2600K but slower than the OC 2600K. The OC 1500X beat out the OC 2600K. All other Rzyen chips handily beat out the 2600K both stock and OC versions. So when you say the 2600K beats out the Ryzen, I have to ask what are you smoking?
Well, there's spam egg sausage and spam, that's not got much spam in it.
Well his source doesn't say what he says they do. First of all the videos don't look at Ryzen performance specifically but do compare Intel CPUs with other ones. They only briefly mention Ryzen; however, in both it clearly shows Ryzen beating out the Intel 2600K he mentions. So he either can't read a graph or is flat out lying.
Well, there's spam egg sausage and spam, that's not got much spam in it.
By that logic, the 2600K has better IPC than a 7700K.
http://www.gamersnexus.net/guides/2867-intel-i7-2600k-2017-benchmark-vs-7700k-1700-more/page-3
That's bad logic.
Since the normal Ryzen line has cheaper CPUs closer to 3.0 Ghz, while the X models start around 3.5.
Maybe the non-X 16-core will actually hit the $850 price point I heard so much about.
The i7-2600k in that article ( http://www.gamersnexus.net/gui... ) is quite a bit behind the new Ryzen CPUs for most tests, even though they have it running at 4.7 GHz vs. the Ryzen's 4 GHz with 2933 MHz RAM. The 2600k is only really winning for certain games.
Ryzen overclocking has improved a bit since that article was published, but you'll still be hard-pressed to get a Ryzen CPU much past 4 GHz. What does matter a lot more is running your memory at a higher speed. Since the infinity fabric interconnect is tied to the memory clock, any time a core needs to reach out over that interconnect to talk to another core, you're going to wait on the infinity fabric. Increasing memory speed has a significant effect on Ryzen in synthetic CPU tests and in many games. Additionally, there have been some game updates for Ryzen that have massively improved performance, and some Windows scheduler updates (I think Windows 10 only) that have yielded a moderate improvement.
I myself have a i7-2600k in my main system. I intend to build a Ryzen (or Threadripper) system later this year.
Since gaming is not very well threaded, and in clock-limited situations these won't be any better than Ryzen. Intel still wins out for performance there, and these are expensive processors to be using in applications which won't use all the cores. Oh well, I guess some folks may just want them for bragging rights?
William George
Threadripper: It has the electrolytes processes crave
The Daddy casts sleep on the Baby. The Baby resists!
Ubuntu server with KVM virtualization (including a Win10 gaming VM) actually.
I do admit I was hoping to be a bit further away from 800 bucks on the 12 core, though. Oh well, gotta wait and see the prices in Switzerland first anyway.
If you stream high quality video encoded with software rather than the dedicated hardware, you need all the cores you can get. Encoding video in software still gives a much higher image quality than the hardware accelerated encoding.
Someone made a performance testing video showing Ryzen greatly outperforming an i7 when simultaneously encoding video and playing the game.
the public. Intel's current prices are highway robbery.
evil greedy executives sure like to cheat and steal
The i3 7350K CPU can be overclocked to at least 5Ghz, giving fantastic single-threaded performance, comparable to an overclocked i7. Many tasks still care about single-threaded performance more than multiple cores.
See subject: Increase cpu core count @ hardware level (OS can use it for starters ala this in Windows for example):
Windows Registry Editor Version 5.00
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Executive]
"AdditionalCriticalWorkerThreads"=dword:00000008
"AdditionalDelayedWorkerThreads"=dword:00000008
* I.E. - How much extra cores will help BEYOND today's CPUs for the OPERATING SYSTEM itself (in Critical Worker Threads) in juggling threads in itself & for other processes (in Delayed Worker Threads) per https://msdn.microsoft.com/en-us/library/cc615012(v=bts.10).aspx/
Here I use 8 for an Intel Core I7 as shown above (both in 1st a 920 & currently a 4790k, since they're quad core (& hyperthreaded) & it was lesser based on physical core count of earlier systems I had (this setting has been around since, iirc, Win2k (correct me IF I am off/wrong - it's been SO long since then)...
(Those are settings in WINDOWS you can adjust to take advantage of added cores as you upgrade to CPUs w/ more cores, for example).
ANYTHING/EVERYTHING, in theory, gains there alone (less "process scheduler thrashing" in other words) - I don't care so much about applications/programs (they are probably written to their practical limits anyhow as to what threadwork will gain them) but again, MORE about how the OS will utilize them (per the 2 TUNABLE PARAMETERS in the .reg file I note above as a way to REALLY use the extra cores, almost guaranteed - Windows allows it, not sure of other OS like *NIX based ones).
APK
P.S.=> The rest will be done @ compiler level (already good, only depends on HOW you can leverage it OR if internal-to-program itself datasets allow for it - not all do) & it's always that way, pretty much - hardware 1st, software catches up (& it does, mostly inefficiently @ 1st, sucking up the CPU cycles/efficiencies gained)... apk
...a 20,000 BTU air conditioner, to offset the heat created by AMD product.
Let me know when there's quality Android product in this class: Reaper DAW, Steven Slate Drums, thousands of VSTs, Prominy Rock Bass. All running on multiple 4K monitors.
Feel free to drop me a line when.
You really need to recalibrate your "is fucked" dowsing rod.
After a decade of living on cream and sunshine, summer vacation is over, and Intel will soon have to buckle down and earn good grades, obtained through long hours of hard study.
Heard on beaches the world over when governments shorten their unemployment insurance entitlement periods: "Shit! We're fucked! Now we'll all have to get real jobs."
Welcome back to how everyone else lives.
On the one hand, it will take a while for Intel to recover its former work ethic. On the other hand, they're well rested—and surrounded by three decades worth of motivational trophy cases (lately somewhat dusty) for Best of Breed in the 800-lb gorilla division.
Yeah, that sure sounds fucked, doesn't it?
The bottleneck for most problems isn't CPU cycles/second, it's the bandwidth of getting data to/from those CPUs. Adding CPUs does nothing to improve performance unless you also give it a much wider I/O path to memory. Adding more high speed cache, on the other hand, may help more than adding CPUs.
I've abandoned my search for truth; now I'm just looking for some useful delusions.
You lying shill.
Change your name to BillyGoatse and then kill yourself
I agree. I use every one of my 8 cores on my AMD.
1 - Vulkan API supports ALL CORE utilization. That's the whole point of Vulkan.
2 - Hosting a VM (for all your work stuff, for example) means I can dedicate 4-6 cores and have both my systems completely independent for all practical purposes. My work SQL instance won't freeze my gaming. Working from home, that's a huge benefit for me to not need to buy two machines or to switch back-and-forth. With RAM being super cheap (32 GB for $100!) I can dedicate half to the machine as well and still have a whopping 16 GB for myself.
3 - Video encoding for streaming. (Although try 4K video encoding without NVENC / GPU and enjoy your 2 FPS.) Other things. It's ALWAYS better to have "too many" cores than "too few." I can spin up KDENLIVE (a great FOSS non-linear video editor!) to render a video, while I'm still playing a game and recording it, or watching 4K videos.
It's really nice to be able to have your computer respond as fast as, your ability to think and queue up operations. (SSDs are a night-and-day difference. I would never go without one.) And at least 8 cores really helps that become a reality.
... don't let their be "processor hang bugs" like FMA3 bug.
I need this processor to replace my 12 year old system !
Ryzen CPU hangs/causes kernel panics and MCE errors, most easily reproducible on Linux
https://community.amd.com/thread/215773?start=180&tstart=0
I hope they can finally fix it. Newest posts (as of today) claim people who RMA'd their CPU and got replacement experience improvement.
Does AMD have anything resembling Intel's AMT malware ?
For me, one of the most crucial parameters when choosing my next CPU is to get one without a backdoor in the chipset (after the vulnerabilities discovered in Intel Management Engine). Intel is no longer a player. AMD has had similar "features" in the past. Does anyone know if they have carved that crap out of this new series of CPU's?
There won't be any Alienware setups that are CPU bound in gaming. These will be attached to either high resolution or high frame rate monitors (or both) and will be GPU bound. And you'll also get the benefit of multiple cores for everything else.
Oh yeah, using multiple VMs or doing video editing are great cases for higher core count processors! I am under the impression that Alienware's target market is gaming, though, and even though newer APIs like Vulkan are making the move toward multiple cores it is still the case that today's games tend to favor higher clock speed over more cores (once you have 4 or maybe 6 cores). Have a 16 or even 12-core dedicated gaming system is a waste, you'd be better off spending the extra money on a more powerful GPU.
William George
Now that is a fair point, if you are using CPU based video recording / streaming while gaming. I personally prefer GPU-based streaming myself, but I have heard some folks say that CPU based can give better looking results with a sufficiently powerful system. I wonder if programs like OBS can really utilize that many cores effectively, though? If you have a 12 or 16 core processor and are playing a modern game that needs 4 cores (a good average) then you have 8-12 cores left over for CPU encoding. I have not seen tests looking at whether OBS and similar applications can really utilize that many cores effectively, but it would be interesting to know.
William George