Are we actually ever going to see IPC improvements ever, or is this it now?
>>61028303
Wait for Zen 2(tm).
Get Ryzen, Intel already lost in IPC.
>>61028361
At best it'll match Sandy Bridge.
>>61028411
Then how come that the i7-7820X, which has the same clock rate, price and number of cores as the R7 1800X, is so much faster everywhere (even in Ryzen's beloved Cinebench)?
>>61028303
BINGBUS
>>61028303
Never, AMD is finished and bankrupt.
>>61028537
That is about the difference between Ryzen and Intel IPC, 10%.
>>61028537
Doesn't the 7820k clock higher than the 1800X because of the new turbo functions?
>>61028537
>single thread
>>61028830
>what is IPC?
Get off this board.
>>61028849
>what is turbo clock
>>61028537
It's probably RAM bottlenecked.
>>61028535
Can't tell if LARPing or retarded.
>>61028933
>he thinks AMD won't pull yet another blunder
>>61028961
They're not gonna go backwards.
Thanks for answering my question though. It was too much to assume someone on this board would be LARPing.
>>61029021
>backwards
Since when has AMD produced anything to match the single core performance of Sandy?
Daily reminder that post-32nm dies suffer from significant transistor decay. So, stick with your Sandy Bridge and Piledriver chips (and pre-GCC amd gpu's and whatever Nvidia had before their 28nm chips).
>>61029075
2017, they released their new Ryzen chips, which surpassed sandy bridge in single core performance.
>>61029075
Multicore since FX, single core since Ryzen you IMG_XXXX imbecile
>>61029112
>>61029113
>he actually believes this
Cute.
>>61029075
are you live under rock?
>>61028303
Difference in ipc between sandy and skylake/kabylake is about 20-25%
>>61029138
So this... is the power of intel
Ryzen and Kabylake are very close in terms of IPC. Don't know why anyone is even bringing up Sandy Vag.
>>61029171
>>61029113
>>61029112
>claim AMD/Intel making progress in single core performance
>7700K shits on everything in AMD's arsenal in games
>clock for clock almost zero progress by Intel
Explain.
>>61028537
13.5% in single core seems about right for the ipc of each.
>>61029112
>which surpassed sandy bridge in single core performance.
Clock-to-clock, no it doesn't, not in real life applications. SHOC 1.004, ARMA2, with no gpu bottleneck, [email protected] and [email protected] (so, both with stock), Zen loses with minimum fps, average being identical. DDR3@1600MHz CL9 and DDR4@3200MHz CL10 (Samsung B-die), P8P67(B3) Deluxe and Taichi X370 (with latest uefi with 1.0.0.6 AGESA).
>>61029311
>He thinks some shitty military simulator is a good IPC benchmark
Time to fuck off back to /v/, kid.
>>61028537
>1800X xfr 4.1ghz
>7820X max turbo 4.5ghz
same clock speed
>>61029322
>synthetic benchmarks matter more than real life ones
Fuck off.
>>61029311
so this is the power of intel shill
>>61029401
>>61029413
>b-but muh ARMA ..
>>61028849
IPC doesn't preclude multiple processors.
Also Turbo Clock 3.0 makes this test disingenuous and you know it.
>>61029311
And this with NT6.1 SP1 with 2016 Convenience rollup (and it's prerequisite service stack hotfix), June and September 2016 rollups, with no other hotfixes. GTX980 on both, 350.12 driver version.
There might be larger difference on Windows '10', but it is unlikely, because almost all tests give better results on 7/Server 2008 with Zen. And besides: not likely scenario, since no-one uses said OS.
>>61028303
Instructions Per Clock
>entire thread is absolute performance
Even Tryzen is better than a sandy with clock speeds equalized
>>61029413
Funny results, I have over 10fps higher minimum with 2500k@stock, DDR3@1600MHz CL9 and 20fps higher average on said game. Have not tested with 1700x.
>>61029536
And for the record: 5820k@stock, DDR4@2133MHZ (same Samsung B-die), Asus X99-M WS: only two frames higher minimum and no difference in average compared to 2500k@stock.
>>61029413
so this is the power of amd shill.
Don't get hit by a car in your bathroom
>>61029311
>gaymez are real world applications
EVERY
FUCKING
TIME
>>61029413
>>61029740
You're both retards for posting MUH OC'ED GAYMAN benchmarks in an IPC thread.
>>61029777
You're retarded for even opening this thread
>>61029081
Daily reminder that you're full of shit kek
>>61029777
This is what happens when shills from both sides invade a legitimately good thread.
>>61029413
Nice cherry picked bench
>>61028849
>instructions per clock
>don't show how they perform at the same clockspeeds
>>61029311
>mentioning ARMA
>one of the shittiest optimized games of all time
>>61029170
It's actually about 30%, which is why the i5 7600K can match the i7 2600K in a lot of multithreaded applications.
STOP STOP STOP!111!!!! EVERYONE IS A CONtent CREAtor now! gaymen hahaa
>>61028303
When they have to. If they are smart they are pushing R&D right now. They dont have anything waiting up their sleeve or we would have seen it in enterprise where the fucking chips are like 6k easily.
>>61029232
>7700K shits on everything in AMD's arsenal in games
>5fps when its running at over 100 is shitting on it
gaymers everyone
>>61029812
this was a shit thread to begin with (OP)[FGT] and half of neo-g don't even know what ipc means or we wouldn't have 49 repliues to something so obviously wrong.
>>61029960
Then why are you staying in a thread that you think is shitty? Oh wait, you're actually having fun!
>>61029917
Reminder to report shitposters.
>>61029936
>TFW you relies that the gaming industry is the only thing keeping the desktop PC relevant. And content creators just use MACs.
>>61030083
I hate this board
>>61028303
So to actually make an answere op's question like you retards never do:
I think there is still room for improvement but not alot. I wouldn't expect to much IPC wise in the next year's.
The future is multiprocessing. There is a lot more possible in this area which is why everyone does it now. I hope Moore's law continues until 2035 so I can get my perfect robot AI gf before I get old.
>>61030083
Hmm I wonder if putting anything above 4 cores at insane pricing for more than half a decade had anything to do with that.
>>61028303
still riding out my SB rig now, waiting for the actual SB 2
>>61030111
GTFO then
https://chiru.no/u/average_intel_owner.webm