Xeon Platinum and Gold CB15 benchmarks

[leadeater]

Just now, MageTank said:

Intel ran out of obscure nomenclatures, so they steal Pokemon's? Do they think we are going to buy these Xeon's just to "Catch 'Em All"?

Xeon Leaf Green for the low power models?

[MageTank]

Just now, leadeater said:

Xeon Leaf Green for the low power models?

I'd totally buy it, just for the name alone, lol. If AMD counters with "Ryzen Naples Fire Red" and makes it run hotter, they'd win the internet.

[The Benjamins]

2 minutes ago, MageTank said:

Intel ran out of obscure nomenclatures, so they steal Pokemon's? Do they think we are going to buy these Xeon's just to "Catch 'Em All"?

1 in 40,000 while be a shiny, They will be the best out there.

[NumLock21]

4 hours ago, tlink said:

what are those holes in the heat spreader? is it for prototyping or will it be in the final result?

There are different answers to that

1. For regulating pressure

2. To let the gasses escape during the manufacturing process

3. A pre-made hole, so it can made it into a key-chain, when the cpu is broken.

It's found most high-end cpus.

[leadeater]

Just now, MageTank said:

I'd totally buy it, just for the name alone, lol. If AMD counters with "Ryzen Naples Fire Red" and makes it run hotter, they'd win the internet.

"After 3400 hours of CineBench Intel i7 6900K evolves to Intel i9 7900X"

[leadeater]

3 minutes ago, MageTank said:

I'd totally buy it, just for the name alone, lol. If AMD counters with "Ryzen Naples Fire Red" and makes it run hotter, they'd win the internet.

Maybe AMD is the Digimon of CPUs? "Ryzen warp digivolves to EYPC"

 

Sorry had to, I'll stop with the lame ass jokes now 

[grimreeper132]

the name schemes stupid, and makes no sense, I feel that a naming scheme should pretty much be

 

WXXYYZ

 

W- gen number, e.g. 7 for kaby lakes gen the 7th gen

XX- number of cores

YY- Base Clock speed

Z- this one would be simple enough it would consist of a letter

K- unlocked

L- Laptop CPU

N- not unlocked

S- Server xeon CPU

 

There would be no issues with this far as I can see why the hell can't the do this. you could still call them i5, i3, i7, i9, xeons etc. but this way you could look at the name get basic specs and not get wrapped up with the tiers which just makes this entire thing pointless

[leadeater]

Just now, grimreeper132 said:

the name schemes stupid, and makes no sense, I feel that a naming scheme should pretty much be

 

WXXYYZ

 

W- gen number, e.g. 7 for kaby lakes gen the 7th gen

XX- number of cores

YY- Base Clock speed

Z- this one would be simple enough it would consist of a letter

K- unlocked

L- Laptop CPU

N- not unlocked

S- Server xeon CPU

 

There would be no issues with this far as I can see why the hell can't the do this. you could still call them i5, i3, i7, i9, xeons etc. but this way you could look at the name get basic specs and not get wrapped up with the tiers which just makes this entire thing pointless

The old Xeon naming scheme was that, it was actually very good. Why the hell they change it I have no idea.

 

https://www.intel.com/content/www/us/en/processors/processor-numbers-data-center.html

[MageTank]

Just now, leadeater said:

Maybe AMD is the Digimon of CPUs? "Ryzen warp digivolves to EYPC"

 

Sorry had to, I'll stop with the lame ass jokes now 

The only joke in this thread is Intel, and their ever-evolving (ha) product names. 

 

Wonder what they plan to do in the future? Surely they have to one-up themselves. How else can they keep their customers confused?

[grimreeper132]

Just now, leadeater said:

The old Xeon naming scheme was that, it was actually very good. Why the hell they change it I have no idea.

exactly it was simple to follow, and made sense. There is also no clashes between gens due to the gen number and you don't get confused between the two different E5 2669 v3 (18 core) and v4 (22 core) as easily as they have entirely different names instead of one number at the end of the name, which half the time is left out as people don't think it's important so you have no clue what fucking CPU they are talking about

 

(if you haven't got it by now, intels and AMDs naming schemes really annoy me)

[Flavio hc 16]

my point is: who is going to buy the 28 cores instead of the 24 cores? The difference in clock is so high that the 24 core has 64,8 ghz of "single core frequency" (I hope you get my point), meanwhile, the 28 core has 58.8. Even if your workload is highly multithreaded,  and of course it will be if you are shopping for those cpus, the 24cores one will be still faster. Unless there is a problem of tdp and/or price

[The Benjamins]

3 minutes ago, Flavio hc 16 said:

my point is: who is going to buy the 28 cores instead of the 24 cores? The difference in clock is so high that the 24 core has 64,8 ghz of "single core frequency" (I hope you get my point), meanwhile, the 28 core has 58.8. Even if your workload is highly multithreaded,  and of course it will be if you are shopping for those cpus, the 24cores one will be still faster. Unless there is a problem of tdp and/or price

one reason is VM farms, they sell per core. but in all I agree.

[leadeater]

2 minutes ago, Flavio hc 16 said:

my point is: who is going to buy the 28 cores instead of the 24 cores? The difference in clock is so high that the 24 core has 64,8 ghz of "single core frequency" (I hope you get my point), meanwhile, the 28 core has 58.8. Even if your workload is highly multithreaded,  and of course it will be if you are shopping for those cpus, the 24cores one will be still faster. Unless there is a problem of tdp and/or price

The higher core count is useful in virtual hosts where pCPU:vCPU ratio is important and not just the overall physical CPU performance.

[Flavio hc 16]

3 minutes ago, The Benjamins said:

one reason is VM farms, they sell per core. but in all I agree.

 

2 minutes ago, leadeater said:

The higher core count is useful in virtual hosts where pCPU:vCPU ratio is important and not just the overall physical CPU performance.

thanks for the explenation

[Curufinwe_wins]

48 minutes ago, Flavio hc 16 said:

my point is: who is going to buy the 28 cores instead of the 24 cores? The difference in clock is so high that the 24 core has 64,8 ghz of "single core frequency" (I hope you get my point), meanwhile, the 28 core has 58.8. Even if your workload is highly multithreaded,  and of course it will be if you are shopping for those cpus, the 24cores one will be still faster. Unless there is a problem of tdp and/or price

Also these are base clocks.... the 28C one outside of thermal limits, hits 2.8 all core turbo.

 

The 24C one hits 3.3.

 

So about the same throughput assuming linear scaling. 

[Coaxialgamer]

6 hours ago, tlink said:

what are those holes in the heat spreader? is it for prototyping or will it be in the final result?

From what i've seen that's a common thing in large server chips . Might be for the soldering process . Pentium 4 also had some.

[sazrocks]

Do we have any price rumors yet?

[grimreeper132]

7 minutes ago, sazrocks said:

Do we have any price rumors yet?

don't think so. But I don't think we have an idea on when release is either so give it time

[laminutederire]

1 hour ago, The Benjamins said:

1 in 40,000 while be a shiny, They will be the best out there.

But it won't be soldered so people will delid them to make jewelry in order to make profits off of them

[mach]

Stop using cringy names ffs intel and amd

These are professional grade products just give them numbers the users will find out what they need anyway

[Taf the Ghost]

40 minutes ago, mach said:

Stop using cringy names ffs intel and amd

These are professional grade products just give them numbers the users will find out what they need anyway

Most of the buyers of the Tech like this are in their 30s to 40s, so they grew up with this stuff. Just wait another decade for "Radical" to make a return.