Enigma2 streaming to the internet WITHOUT a PC

[kunmun23]

1.3 GHz only means that the CPU can do 1.3GHz cycles BUT not every cycle is an instruction. Division for instance takes X cycles on CPU A and Y cycles on CPU B. Or CPU B does not have the instruction Z so it must be done by a subroutine which is a lot slower.

YOU CANNOT COMPARE CPUs simply by comparing the system clock!!!!!!!!!

ciao

Totally agree that CPUs can't be compared solely by system clock, but given a comparable instruction set (i.e. both the systems takes x cycles for a divide operation) you can compare CPU speeds. For instance two CPUs on a x86 architecture using let's say, intel instruction set, one at 1.4GHz and one at 2GHz. The 2GHz would surely be faster in processing the same piece of information than the 1.4GHz.

CPU speed isn't a pure marketing hogwash. I do agree that it should not be the only metric to look at when comparing system performance.

Anyways, coming back to the point :

Are you saying that a 1.4GHz ARM Cortex A7 quad core android dongle does not have enough processing capability to transcode on the fly ??


Would love to hear views from some other experts too.

[pac2000]

MIPS/ARM Processors Are RISC (Reduced Instruction Set Computing) Based Whereas x86/x64 Are CISC(Complex Instruction Set Computing)Based... Trial Is 100% Correct That You Cannot Under Any Circumstances Compare The 2 Based On Clock Speed Or Amount Of Cores...

Here's A Z80 Instruction LDDR

This Is It In Another Form (The Routine In Z80 If The LDDR Instruction Didn't Exist)

Loop:
LD (DE),(HL)
DEC DE
DEC HL
DEC BC
LD A,B
JR NZ,Loop
LD A,C
JR NZ,Loop

The LDDR Instruction For Starters Only Processes 1 Instruction... It Doesn't Need To Pull More Instructions In From Memory And Increment The Program Counter For Each Of The Instructions Above... This Saves Huge Amount Of Time... This Routine Is Hard Coded In The Processor And Is One Of Many In A CISC Processor... With RISC You Must Write These Routines, Which Are Very Slow By Comparison, And Take Up More Space In Memory... And In Fact Even For Some Instructions In That Routine Above In RISC Code Would Require Several Instructions Per CISC Instruction...

The LDDR Instruction If You Haven't Guessed Is Used For Shifting Blocks Of Memory From One Location To Another...

You Can't Honestly Think A Cheap ARM Processor Can Compare To An Expensive Intel/AMD... And I Haven't Even Gone Into Pre-Fetching And Pre-Processing That Modern CISC Architecture Perform...

[Trial]

Hi,

For instance two CPUs on a x86 architecture using let's say, intel instruction set, one at 1.4GHz and one at 2GHz. The 2GHz would surely be faster in processing the same piece of information than the 1.4GHz.

here you must be careful. If you compare 1st generation cores with last generation haswell the later will be more than twice as fast with the same clock. I have an old dual core with 2*3.4GHz which will be easily beaten by current Atoms with half the clock. So clock says simply almost nothing. It is only a vague indicator for the speed.

ciao