i5-3570K is 3.4 GHz, came out in 2012
i5-6600K is 3.5 GHz, came out in 2015
that's hardly any difference in speed, why is this allowed?
>>55104355
Because literally jews
>>55104355
clock speed is a bad measure of performance. IPC exists, and if two processors have the same clock speed but one has twice as much IPC, then it is two times faster....
clock speed x instructions per clock / watts = true measure of a cpu
>>55104382
everything you said is true, but IPC gains have not been very impressive from 2012-now
>>55104355
Why didn't you mention the 2.7ghz i5 6400?
>>55104355
i got the ivy
how much perf am I missing out for not upgrading to skyjoo?
>>55104433
In vector ops they have.
Serial integer performance isn't going to make any big gains like we used to see. Diminishing returns is a law.
>>55104355
Fortunately AMD saved us with the 9590 and 4.7GHz (5GHz Turbo!!!) out of the box, making it the best CPU ever made.
THANK YOU BASED AMD
The Ivy Bridge CPUs are better because they can be clocked higher.
http://hwbot.org/benchmark/superpi_-_1m/rankings?#start=0#interval=20
>>55104382
This is a large part of the reason why, but I think another aspect of it is that desktop PC sales don't make up the majority of the market anymore, laptops do and Intel knows that. They also know that on a laptop, efficiency and volume are more of a premium than raw performance. Since their desktop and laptop CPUs are more or less the same, they've been putting their efforts into improving things like performance per watt or physical cpu size. These improvements do carry over to the desktop market, but they're effectively useless there. Laptops on the other hand, have improved significantly over the last few years.
Nvidia is also doing the same thing with their GPUs. The 9xxM series GPUs were amazing for their performance per watt.
You can also see this trend in the desktop market when you consider that a 1080 and an i7 can be safely powered with less than 600W.
Hell, my 3770k is still fast as hell, 2 generations old even
>>55104458
I was just comparing the one I have now
>>55104382
where is IPC listed here?
http://ark.intel.com/products/65520/Intel-Core-i5-3570K-Processor-6M-Cache-up-to-3_80-GHz
>>55104816
IPC is meaningless. A CPU with triple the IPC of Skylake would be garbage if it can't get a clock speed above 1 GHz.
>>55104826
This. Raw clock speed is really the only indicator that matters when comparing CPUs.
>>55104816
You can't put a single number on IPC because it varies between different kinds of processing workload. The simplest and fairly accurate way is to look at Passmark multi-threaded and single-threaded scores.
>>55104355
It is actually worse: Pentium 4 was 3.8GHz in 2005.
Intelluminati NWO Barack Hussein Abudalla Obama are consipirizing to steal our Gigahertz!
>>55104826
Retard like you are probably delusional poorfags clinging on to their Core2Duo full voltage Thinkpads.
>>55104845
>CPU1 with 3GHz clock
>CPU2 with 1GHz clock and 6 times higher IPC than CPU1
Hurr durr CPU1 is faster
Meanwhile other companies are selling 48 core CPUs that beat $2000 xeons in some applications for $800.
This thing is made on an old 28nm fab by a company smaller than AMD and it was their first time making a CPU.
The next 16nm generation are going to be more painful for intel.
>>55104919
Fuck off shill. "IPC" is just a meme to sell new CPUs.
>>55104974
So what's the reason the single core performance of my Xeon e3-1231v3 is higher than the single core performance of Pentium 4 Extreme Edition 3.46
>>55104992
It isn't. Rigged benchmarking programs paid for by Intel to help sell CPUs prove nothing except that you're gullible and fell into their trap.
>>55104884
Except the i5-6600K will, under single-core loads, go up to 3.9GHz w/o overclocking
>>55104826
>>55104845
A 1.7MHz 6502 runs code faster than a 3.5MHz Z80
Clock speeds was a Zilog scam long before IPC was an Intel scam
...and Zilog was founded by former Intel employees
>>55104992
>>55105006
If you want a benchmark, install POV-Ray and run scenes/advanced/benchmark/
It's not made by Intel, there's fuck-all SIMD stuff, and the Xeon will still beat the P4
Plus when you're done you can make fancy shiny shit like chrome balls floating above chequerboard floors
>>55104791
why not compare two processors of the same speed?
moron
>>55105087
Because it's more fun to watch a 1GHz CPU kick the shit out of a 3GHz CPU
>>55104926
141w idle power draw. Around the clock.
It only beats a low power Xeon D in a few metrics, in everything else its markedly worse in perf/watt. The ARM based ThunderX draws up to 240w. The Xeon D with 65w TDP draws just 124 peak, its peak draw under full load is lower than the ThunderX's idle draw.
Its a horrible, absolutely atrocious platform. Matter of factly its god fucking awful. ARM still doesn't have a shot at competing with X86 in performance.
>>55104599
Same, though I got it running on a waterloop @4.8GHz.
>>55105006
Oh ho ho ho. Nice justification, poorfag. Now show us how every single benchmark was paid off.
my 3570k @ 4.5 works fine after all these years
dunno why people bother to upgrade
>>55104926
I can program a $200 FPGA to beat xeons in some applications.
>>55104355
>what is ebay and Craigslist.
Every fucking one of you are normies. And no this isn't fucking b8. I just got an I7 4790k for 200$ brand new never opened off Craigslist. And evga motherboard for 120$. Quit your bitching about Intel being jews and hunt for deals. All you nigger want is shit given to you cheaply
You got to options
>1. Stop being poor
>2. Look for deals
>>55105561
>you got 2 options.
Typo
>£140 i5 750 when new 2.67ghz
>£140 i5 6400 today 2.7ghz
30mhz is not enough of an improvement for 6years of progress.
>>55105773
And that new CPU will be 4x faster
chips are bigger
because of integrated GPUs and other stuff
speed of light limit still exists
therefore you cant get higher than like 6Ghz without breaking the laws of physics
only if you stacked the CPU (a cube instead of 2D rectangle/square)
>>55106216
>2.67 x 4 = 2.7
The Burgerclap education system, ladies and gentlemen.
>>55106323
Benchmarks show me to be correct. IPC matters more than you fags will admit.
>>55105773
https://cpubenchmark.net/cpu.php?cpu=Intel+Core+i5-750+%40+2.67GHz
https://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i5-6400+%40+2.70GHz
About 75% faster.
>>55106481
But the speed is the same! Jewish tricks to fool the goyim
