32C/64T AMD Ryzen Threadripper Cinebench scores surface
Quote:
And in a CPU-Z screenshot the CPU is described as the 'Ryzen Threadripper 2990X'.
Read more.
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Difference this would make to my non-linear analysis and FEM work-flow. High. Chances of work upgrading my i5-2500 to anything with more than 4 cores? Low. Sigh... :rolleyes:
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Quote:
Originally Posted by
ik9000
Difference this would make to my non-linear analysis and FEM work-flow. High. Chances of work upgrading my i5-2500 to anything with more than 4 cores? Low. Sigh... :rolleyes:
have you priced a Ryzen 7 1700? or 2700 maybe? The 8 core, 16 thread beast might be in reach.
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Quote:
Originally Posted by
ik9000
Difference this would make to my non-linear analysis and FEM work-flow. High....
Can you "borrow" a high core count machine to run a workload on and demostrate to management how much of your time they could be saving by investing in faster tech? Sounds to me like there's a business case to be made there...
At first sight 32 cores @ a 3.4GHz base clock looked very impressive, but then I realised that a 250W TDP is 62.5W per die. Once you take into account the fact that 2 dies have their IMCs disabled and threadripper is meant to get cherry picked silicon, it's actually a bit dull that they're running at 65W Ryzen clocks....
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Quote:
Originally Posted by
ik9000
Difference this would make to my non-linear analysis and FEM work-flow. High. Chances of work upgrading my i5-2500 to anything with more than 4 cores? Low. Sigh... :rolleyes:
You gotta make the fiscle argument, its not that it will make your job 16 time easier, it will make you 16 times cheeper! And still the manager will think: hey it comes out of my budget and my bonus is based on being under... Damn suits!
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Quote:
Originally Posted by
scaryjim
At first sight 32 cores @ a 3.4GHz base clock looked very impressive, but then I realised that a 250W TDP is 62.5W per die. Once you take into account the fact that 2 dies have their IMCs disabled and threadripper is meant to get cherry picked silicon, it's actually a bit dull that they're running at 65W Ryzen clocks....
Maybe these ones are from the same batch as the 65W Ryzen chips, instead of being the cream-of-the-crop the 2-chip Threadrippers are made from?
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Quote:
Originally Posted by
afiretruck
Maybe these ones are from the same batch as the 65W Ryzen chips, instead of being the cream-of-the-crop the 2-chip Threadrippers are made from?
Hmmm, I can't see them wanting to damage the premium image they've been trying to cultivate for Threadripper, and tbf the 65W 8 core Ryzen (2700) bases at 3.2GHz. The really interesting thing will be when the proper reviews come in and we start to see what they can maintain under PB2 - that 3.2GHz R7 2700 actually maintained > 3.4GHz under load, so if the Threadrippers have a 3.4GHz base but can actually maintain 3.6GHz+.... :surprised:
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Tbf they are cinebenching at 4ghz which appears to be pretty stable to run the benchies. So hopefully although they might be running at 65w per die they are also able to hit better clockspeeds
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Quote:
Originally Posted by
scaryjim
Hmmm, I can't see them wanting to damage the premium image they've been trying to cultivate for Threadripper, and tbf the 65W 8 core Ryzen (2700) bases at 3.2GHz. The really interesting thing will be when the proper reviews come in and we start to see what they can maintain under PB2 - that 3.2GHz R7 2700 actually maintained > 3.4GHz under load, so if the Threadrippers have a 3.4GHz base but can actually maintain 3.6GHz+.... :surprised:
Maintaining 3.6GHz would be really impressive under full load! Thinking about it, I imagine they'd want to keep the cream-of-the-crop chips to allow maintaining higher boosts (4 GHz+?) under lower loads.
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Quote:
Originally Posted by
ik9000
Difference this would make to my non-linear analysis and FEM work-flow. High. Chances of work upgrading my i5-2500 to anything with more than 4 cores? Low. Sigh... :rolleyes:
The Ryzen 5 1600 is under £150 now.
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Imagine if they had an ac unit strapped to that thing
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Assuming perfect scaling, 4x the 2700 cinebench score is 6252. That's the same ballpark as the numbers in the article, so it can't be running an all core boost too far from the 2700 figure.
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Y'know what's really funny though.... Anandtech just did an article showing how that for 4K gaming, AMD's Ryzen 3 is pretty much as good as the top tier CPUs. So save a few hundred quid and throw it at the GPU was their advice.
But part of me really wants a threadripper.
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Does it need a $5000 dollar cooling solution like the recent Intel did? If i recall correctly the Intel rig scored 7400-ish at 5ghz.
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Quote:
Originally Posted by
ohmaheid
Does it need a $5000 dollar cooling solution like the recent Intel did? If i recall correctly the Intel rig scored 7400-ish at 5ghz.
It's 250 watts and a bigger socket. That's pretty easy to do and with decent air cooling I would imagine not a problem
Re: 32C/64T AMD Ryzen Threadripper Cinebench scores surface
Quote:
Originally Posted by
3dcandy
It's 250 watts and a bigger socket. That's pretty easy to do and with decent air cooling I would imagine not a problem
Coolermaster have already created (and has been shown) an air cooler designed for Threadripper 2 but I doubt it would be able to sustain an all core clock of 4+ GHz.