• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Skylake-X Lineup Leaked: i9-7980XE 18 Core Flagship Processor

Soldato
Joined
18 Oct 2012
Posts
4,148
Location
Oxfordshire
the 7820x only needs to clock at 3.5ghz to be on par with a 4ghz 1800x, Intel has already said the 4.3 boost is an all core boost, and 4.5 is for two cores.

given skylakes faster architecture and higher clockspeeds you're paying basically £100 more for quad channel ram, and 20% more performance.

also, skylake x is different than skylake, they've quadrupled the l2 cache and changed some of the prefetching, apparently a 13% ipc gain over regular skylake due to less misses.

Where did you get those figures that Intel only needs to be 3.5Ghz to be on par with 4Ghz AMD?

Where is this magic 20% performance coming from?
 
Associate
Joined
2 May 2017
Posts
535
Read the small print on the slide on the front page. It refers to the maximum dual core frequency.

that's the 4.5ghz boost, the cpu had two boost functionalities.

boost 2.0 boosts ALL cores (it's the same boost as on 7700ks 4790ks etc)

boost 3.0 is for individual cores, whoch has now been upped to two 'best cores'


Intel literially tweeted that 4.3ghz is the all core boost speed.
 
Associate
Joined
2 May 2017
Posts
535
Where did you get those figures that Intel only needs to be 3.5Ghz to be on par with 4Ghz AMD?

Where is this magic 20% performance coming from?


http://www.anandtech.com/show/11464...ng-18core-hcc-silicon-to-consumers-for-1999/3

skylake is already around 10% faster clock for clock than ryzen, it is 400mhz faster out of the box than an 1800x (on all cores)

the l2 cache is a pretty large change in the setup, with the prefetching changes there's less misses which increase how many instructions get handled per clock cycle.

that's why the broadwell 5775c was often faster than a 6700k in games(clocked slower with slower architecture) purely thanks to a larger l4 cache,obviously l2 cache is far more.important, a 4 fold increase is fairly substantial.
 
Soldato
Joined
12 May 2014
Posts
5,238
that's the 4.5ghz boost, the cpu had two boost functionalities.

boost 2.0 boosts ALL cores (it's the same boost as on 7700ks 4790ks etc)

boost 3.0 is for individual cores, whoch has now been upped to two 'best cores'


Intel literially tweeted that 4.3ghz is the all core boost speed.

So your saying that intel said their own slide is wrong?
 
Soldato
Joined
18 Oct 2012
Posts
4,148
Location
Oxfordshire
Someone has been reading the intel slides :p

Was kinda my point. Not a chance what Intel published will stack up. If we go by every other slide they have ever shown in regards to performance boost and what it has been in real world I don't think we are likely to see any performance boost other than when all cores are utilised.
 
Associate
Joined
2 May 2017
Posts
535
Soldato
Joined
13 Mar 2008
Posts
9,638
Location
Ireland
Lol, Arma is a massive outlier and clearly can't properly use RyZen. IPC wise RyZen is 7-10% in tests compared to Kabylake.

Other games that were also massive outliers have needed specific developer patches to address the issue since the games were out long before RyZen and it's new topology.

Plus if you think Intel are going to manage over gigahertz on massive multi core processors using cheap TIM you're in for a surprise in regards to temps.
 
Soldato
Joined
13 Jun 2009
Posts
6,847
only one thread is used, so it's a clock for clock comparison, why's the ryzen up to what? 40% slower?
Because the game was written and performance tested against Intel's architecture? AMD have had nothing to even bother testing against for years so most developers haven't bothered. We've already seen that a simple patch can boost a game's performance with Ryzen by up to 50% (see Rise of the Tomb Raider). This is why it largely depends on what kinds of games you play - if you're always playing the newest titles, Ryzen fares better and will get even better in the next year or two. The 6c/12t version already beats the equivalently priced Core i5 in the majority of AAA games used for benchmark suites right now and we are not going to go back to single-threaded games.
 
Soldato
Joined
29 Jan 2015
Posts
4,904
Location
West Midlands
that's a 4 core 4 thread vs a 6 core 12 thread in multi threaded titles.

so what you're saying is it takes over double the threads for amd to compete?

here you go https://youtu.be/6yh9doMXt3I
a single threaded game, 4ghz 6700k vs 4ghz 1700.

only one thread is used, so it's a clock for clock comparison, why's the ryzen up to what? 40% slower?

facepalm.jpeg

We all know arma is a piece of **** for anything under 5ghz. That engine needs to die.
 
Associate
Joined
2 May 2017
Posts
535
Lol, Arma is a massive outlier and clearly can't properly use RyZen. IPC wise RyZen is 7-10% in tests compared to Kabylake.

Other games that were also massive outliers have needed specific developer patches to address the issue since the games were out long before RyZen and it's new topology.

Plus if you think Intel are going to manage over gigahertz on massive multi core processors using cheap TIM you're in for a surprise in regards to temps.


10% behind clock for clock on average and 10-15% lower clock speeds.

http://hwbot.org/benchmark/cinebench_-_r15/rankings?cores=10#start=0#interval=20

the new cinebench scores show skylake x definitely has about 15% ipc over broadwell e, if you take away the 300mhz difference (approx 100) you're roughly seeing a 15% clock on clock over broadwell e, which means that Intel are correct and skylake x should have higher ipc than skylake s
 
Associate
Joined
2 May 2017
Posts
535
Because the game was written and performance tested against Intel's architecture? AMD have had nothing to even bother testing against for years so most developers haven't bothered. We've already seen that a simple patch can boost a game's performance with Ryzen by up to 50% (see Rise of the Tomb Raider). This is why it largely depends on what kinds of games you play - if you're always playing the newest titles, Ryzen fares better and will get even better in the next year or two. The 6c/12t version already beats the equivalently priced Core i5 in the majority of AAA games used for benchmark suites right now and we are not going to go back to single-threaded games.


the 1800x falls behind sandy bridge more often than not

http://www.gamersnexus.net/hwreview...iew-premiere-blender-fps-benchmarks?showall=1

it's slower than Haswell, broadwell, skylake and kabylake, soon to be coffeelake too.

so ryzen is only 5 gens behind instead of 10 gens behind Intel:)
 
Soldato
Joined
13 Jun 2009
Posts
6,847
the 1800x falls behind sandy bridge more often than not

http://www.gamersnexus.net/hwreview...iew-premiere-blender-fps-benchmarks?showall=1

it's slower than Haswell, broadwell, skylake and kabylake, soon to be coffeelake too.

so ryzen is only 5 gens behind instead of 10 gens behind Intel:)
This is such nonsense. Even using gamers nexus' strongly criticised zero-day benchmark numbers, Ryzen is certainly not behind anything pre-Haswell. Are you looking at a single game benchmark in an entire suite or something? "More often that not" my arse.
 
Back
Top Bottom