AMD France shares Threadripper 2990WX Cinebench score

[HEXUS]

AMD Ryzen Threadripper 2990WX score: 5,099. Intel Core i9-7980XE score: 3,335.

Read more.

[Tabbykatze]

As I said on the OC3D thread:

Considering the 8180 (appears to be at stock) got 4355 on Cinebench r15 multicore and I'd be very surprised if they didn't hit their core to die limit. For the TR2 32 core to get 5099 is very good.

The 8180 delivered 155 points per core whereas the Tr2 32 core delivered 159 points per core. Which is pretty nifty. This means that stock for stock the Tr2 32 core performed 17% better and costs nearly 80% less than the 8180.

Price vs price though, the Tr2 32 core performs 47% better at a similar price bracket as the i9-7980XE (5099/3455) but it did produce 191 points per core. This means that Intel will have to have an around 24-26 core processor performing at 2.6GHz with 4.2GHz boost to match the Tr2 32 core.

This is possible, Intel could pull out all the stops. However price wise i think the richness of value and compute power will rock the top end range of CPU lineups.

Very impressed with this offering. I am looking forward to Intels answer. This is because Intel will have to pull out all the stops or brutalise their prices

[Ozaron]

...or brutalise their prices

Why do you do this to yourself?

[Tabbykatze]

Why do you do this to yourself?

[philehidiot]

<Mandatory_Rant>

After years of hardly any real progress from Intel and constantly taking the smeg out of consumers with apparently pointless socket changes, confusing product lines, etc I'm quite happy to see them get crushed. By constantly holding back on R&D because they were so dominant in the marketplace they're now not in a position to respond properly. We saw their hasty response to Ryzen - if they'd been doing things properly they'd have had a well tested platform ready to deploy knowing they could hold back if Ryzen flopped. But they had to rush out some answer and it was a mess. Now they're having to do what they should have been doing all along which is to compete and innovate. Really, all I'm seeing is them using marketing tricks to make it look like there's progress when it's really quite limited. In a sense they seem to be doing what AMD had to do all those years ago when they were behind on IPC - throw more cores at the problem.

Intel made this bed, now they can lay in it. I look forward to a resurgent AMD. Will I be buying a TR? As much as I'd love the bragging rights, no.

</Anti_Consumer_Rant>

[Corky34]

You do know Intel spent around $13 billion on R&D in 2017 alone, right? I maybe mistaken but i think that's possibly more than AMD's entire revenue during the same period.

[CAT-THE-FIFTH]

I do wonder how the weird memory controller access will manifest itself in the realworld.

[Tabbykatze]

You do know Intel spent around $13 billion on R&D in 2017 alone, right? I maybe mistaken but i think that maybe more than AMD's entire revenue during the same period.

I think i made a comment about this on the Hexus article about that, you can throw money and people at a problem but it doesn't mean that it'll make magic. You also have to remember than AMD managed to erode Intel enough that they are pumping (slowly) into high gear with a budget that is 10% of what Intel spends. So is that quality vs quantity or is this a case of just quoting Intels or AMDs budget is completely arbitrary because of how diverse the spending is.

[mazty]

You do know Intel spent around $13 billion on R&D in 2017 alone, right? I maybe mistaken but i think that's possibly more than AMD's entire revenue during the same period.

Revenue and investment can't be compared side-by-side as that's not how company finances work. But either way, it looks like Intel are dead in the water as they are struggling with 10nm, whereas AMD, by outsourcing the fabrication, is able to use foundries that can do it without issue.

[DanceswithUnix]

You do know Intel spent around $13 billion on R&D in 2017 alone, right? I maybe mistaken but i think that's possibly more than AMD's entire revenue during the same period.

How much of that was on foundry process improvements rather than CPU design?

$13B doesn't buy a lot of fab, but would be really high for CPU design probably paying for about 60000 well paid & funded engineers.

[Corky34]

I think i made a comment about this on the Hexus article about that, you can throw money and people at a problem but it doesn't mean that it'll make magic. You also have to remember than AMD managed to erode Intel enough that they are pumping (slowly) into high gear with a budget that is 10% of what Intel spends. So is that quality vs quantity or is this a case of just quoting Intels or AMDs budget is completely arbitrary because of how diverse the spending is.

True and i wasn't trying to suggest all the money Intel has spent on R&D has been well spent, i was simply trying to point out to philehidiot that Intel has done anything but constantly holding back on R&D.

Revenue and investment can't be compared side-by-side as that's not how company finances work. But either way, it looks like Intel are dead in the water as they are struggling with 10nm, whereas AMD, by outsourcing the fabrication, is able to use foundries that can do it without issue.

I wasn't trying to compare revenue and investment side-by-side, i was trying to put a £13 billion semiconductor R&D spend into context.

How much of that was on foundry process improvements rather than CPU design?

$13B doesn't buy a lot of fab, but would be really high for CPU design probably paying for about 60000 well paid & funded engineers.

I wouldn't know exactly but apparently TSMC spent around $2.6 billion, GloFlo isn't even on the list i looked at.



I wouldn't have thought it included the buying of machinery as that list of who spent more than a billion on semiconductor R&D has 16 companies on it and there doesn't seem to be much correlation between what they spent and what they're known to have done.

[DanceswithUnix]

True and i wasn't trying to suggest all the money Intel has spent on R&D has been well spent, i was simply trying to point out to philehidiot that Intel has done anything but constantly holding back on R&D.

The P4 and Itanium were products of a well funded R&D at Intel, with crosspoint RAM looking like the next big disappointment.

I've said plenty of times int the past, I don't think Intel are holding back, they have a pretty awful history in CPU design balanced by running the best fabs in the business. It was as much luck as judgement that the mobile pentium parts ended up being decent desktop cores, I think this is just Intel being Intel.

I get the impression they have lots of stuff going on behind the scenes though, carbon nanotube research etc.

[Tabbykatze]

I get the impression they have lots of stuff going on behind the scenes though, carbon nanotube research etc.

I agree, I doubt that $13 billion number is just for CPU. it could also be any startup companies they fund for "cool new ideas" which will account for a sizeable amount. I would also not be surprised if a lot of that budget is going on the "what happens after silicon?" question.

[Iota]

I agree, I doubt that $13 billion number is just for CPU.

Correct, it's definitely for much more than that, including bonuses paid out.

https://www.nasdaq.com/article/heres...-2017-cm948915