SkyBoost (topic 4)

Post » Sat May 19, 2012 10:26 pm

What you're not taking into account in that case is that this mod does more then just change those methods, it also inlines a tonne of extraneous function calls that spend more time getting to and from the function the recovering the memory location the function itself was written to return. That's what the core of tesval did and what skyboost extends upon with further stuff like what he's asking us for feedback upon now.

ED: Er sorry wait scratch that, you're not talking about the mod but how the different methods vary.

Yup, exactly, in that link if you look around you will see guys with:
PhII @3.63 GHz:
x87 : 546ms
SSE : 546ms
SSE2: 811ms

and
My results Core i7 stock
X87:2090
SSE:2050
SSE2:1170

This is what to be expected as before FX AMD had disastrous SSE engines, while Intel ones shined. There should be clearly pronounced difference between x87 and SSE2 on Intel and AMD. I think real culprit is that maybe Alex forgot a compiler flag or sth - it could be that simple.
User avatar
lacy lake
 
Posts: 3450
Joined: Sun Dec 31, 2006 12:13 am

Post » Sat May 19, 2012 11:19 pm

Hmm...

I just tried the Geeks3d test one and got..

-----------------------------
Instruction set: x87
Starting matrix multiplication loop...
Elapsed time: 280 ms
Bye!

-----------------------------
Instruction set: SSE
Starting matrix multiplication loop...
Elapsed time: 249 ms
Bye!

-----------------------------
Instruction set: SSE2
Starting matrix multiplication loop...
Elapsed time: 1092 ms
Bye!

So looking at this, I would have to say there is probably something going on with the sse2 compiler that is not allowing AMD systems to actually use sse2.

Edit: Although using Alex's test, the differences aren't as exaggerated..

---------- testing started ----------
testing x87

Starting matrix multiplication loop...
Elapsed time: 2667 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 2777 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 2776 ms
---------- testing finished ----------
User avatar
Marcia Renton
 
Posts: 3563
Joined: Fri Jan 26, 2007 5:15 am

Post » Sat May 19, 2012 8:36 pm

I just ran the test at the site DariusD linked and here is my result (Phenom II x4 @ 4 GHz) :

Instruction set: x87
Starting matrix multiplication loop...
Elapsed time: 484 ms
Bye!

Instruction set: SSE
Starting matrix multiplication loop...
Elapsed time: 483 ms
Bye!

Instruction set: SSE2
Starting matrix multiplication loop...
Elapsed time: 748 ms
Bye!

And this is my test from earlier (Alexander's test):

---------- testing started ----------
testing x87
Starting matrix multiplication loop...
Elapsed time: 2980 ms

testing sse1
Starting matrix multiplication loop...
Elapsed time: 4009 ms

testing sse2
Starting matrix multiplication loop...
Elapsed time: 4025 ms
---------- testing finished ----------

Edit: no background apps, only the cmd window opened
User avatar
Sasha Brown
 
Posts: 3426
Joined: Sat Jan 20, 2007 4:46 pm

Post » Sun May 20, 2012 5:38 am

Intel i7 X980 4.16 GHz

---------- testing started ----------testing x87Starting matrix multiplication loop...Elapsed time: 1903 mstesting sse1Starting matrix multiplication loop...Elapsed time: 1950 mstesting sse2Starting matrix multiplication loop...Elapsed time: 1794 ms---------- testing finished ----------

less background activity (Firefox)
Spoiler
---------- testing started ----------testing x87Starting matrix multiplication loop...Elapsed time: 1888 mstesting sse1Starting matrix multiplication loop...Elapsed time: 1934 mstesting sse2Starting matrix multiplication loop...Elapsed time: 1778 ms---------- testing finished ----------
User avatar
Katie Samuel
 
Posts: 3384
Joined: Tue Oct 10, 2006 5:20 am

Post » Sun May 20, 2012 1:19 am

Does anybody know if the famous "+fullproc" launch option does any good or bad to Skyboost. And btw, if it does any good at all?
User avatar
CArla HOlbert
 
Posts: 3342
Joined: Wed Feb 21, 2007 11:35 pm

Post » Sun May 20, 2012 4:38 am

i7 920 @ 3.990 GHz, hyperthreading disabled, nothing unusual running in the background:

x87
1981ms

sse1
2028ms

sse2
1888ms

Almost identical results on repeated tests.
User avatar
Ludivine Poussineau
 
Posts: 3353
Joined: Fri Mar 30, 2007 2:49 pm

Post » Sun May 20, 2012 9:48 am

write here ur cpu specs and ouput from the test (or just elapsed time for each stage)

Here you are:

---------- testing started ----------testing x87Starting matrix multiplication loop...Elapsed time: 1528 mstesting sse1Starting matrix multiplication loop...Elapsed time: 1592 mstesting sse2Starting matrix multiplication loop...Elapsed time: 1513 ms---------- testing finished ----------
Intel i7-2600k @ 4.6GHz
User avatar
GRAEME
 
Posts: 3363
Joined: Sat May 19, 2007 2:48 am

Post » Sun May 20, 2012 2:46 am

Here my test.

My CPu = Intel Core i5 750 @ 2.67 GHz stock speed

---------- testing started ----------
testing x87

Starting matrix multiplication loop...
Elapsed time: 3525 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 3463 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 3526 ms
---------- testing finished ----------
Press any key to continue . . .
User avatar
emily grieve
 
Posts: 3408
Joined: Thu Jun 22, 2006 11:55 pm

Post » Sat May 19, 2012 8:44 pm

intel i7 q720 (laptop)



---------- testing started ----------
testing x87

Starting matrix multiplication loop...
Elapsed time: 4493 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 4618 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 4383 ms
User avatar
sas
 
Posts: 3435
Joined: Thu Aug 03, 2006 8:40 am

Post » Sun May 20, 2012 6:52 am

i72600 @stock
---------- testing started ----------
testing x87

Starting matrix multiplication loop...
Elapsed time: 2044 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 2122 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 3479 ms
---------- testing finished ----------
User avatar
roxxii lenaghan
 
Posts: 3388
Joined: Wed Jul 05, 2006 11:53 am

Post » Sat May 19, 2012 9:18 pm

Phenom II x4 945 stock 3 ghz

---------- testing started ----------
testing x87

Starting matrix multiplication loop...
Elapsed time: 3962 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 5101 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 4852 ms
---------- testing finished ----------
User avatar
Jennifer Rose
 
Posts: 3432
Joined: Wed Jan 17, 2007 2:54 pm

Post » Sun May 20, 2012 5:03 am

i7 2600k @ 4.7 Ghz

C:\Users\Derko\Documents\fputest>echo off
---------- testing started ----------
testing x87

Starting matrix multiplication loop...
Elapsed time: 1544 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 1591 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 1497 ms
---------- testing finished ----------
Press any key to continue . . .
User avatar
Stephanie Nieves
 
Posts: 3407
Joined: Mon Apr 02, 2007 10:52 pm

Post » Sat May 19, 2012 7:56 pm

i7 990x @ 4.17 Ghz

---------- testing started ----------
testing x87

Starting matrix multiplication loop...
Elapsed time: 1903 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 1934 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 1795 ms
---------- testing finished ----------
Press any key to continue . . .
User avatar
AnDres MeZa
 
Posts: 3349
Joined: Thu Aug 16, 2007 1:39 pm

Post » Sun May 20, 2012 12:42 am

Athlon II x4 640 @ 3.5 GHz

---------- testing started ----------
testing x87

Starting matrix multiplication loop...
Elapsed time: 3447 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 4727 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 4727 ms
---------- testing finished ----------
User avatar
Vicki Gunn
 
Posts: 3397
Joined: Thu Nov 23, 2006 9:59 am

Post » Sun May 20, 2012 9:12 am

Number of cores 4 (max 8)
Number of threads 8 (max 16)
Name Intel Core i7 920
Codename Bloomfield
Specification Intel® Core™ i7 CPU 920 @ 2.67GHz
Package (platform ID) Socket 1366 LGA (0x1)
CPUID 6.A.5
Extended CPUID 6.1A
Core Stepping D0
Technology 45 nm
TDP Limit 130 Watts
Core Speed 1603.6 MHz
Multiplier x FSB 12.0 x 133.6 MHz
Rated Bus speed 3207.2 MHz
Stock frequency 2666 MHz
Instructions sets MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, EM64T, VT-x
-------------------------------------------------------
Northbridge Intel X58 rev. 12
Southbridge Intel 82801JR (ICH10R) rev. 00
Graphic Interface PCI-Express
PCI-E Link Width x16
PCI-E Max Link Width x16
Memory Type DDR3
Memory Size 6144 MBytes
Channels Triple
Memory Frequency 801.8 MHz (2:12)
CAS# latency (CL) 8.0
RAS# to CAS# delay (tRCD) 8
RAS# Precharge (tRP) 8
Cycle Time (tRAS) 24
Row Refresh Cycle Time (tRFC) 88
Command Rate (CR) 1T
Uncore Frequency 3207.2 MHz
------------------------------------------
testing x87

Starting matrix multiplication loop...
Elapsed time: 2808 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 2870 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 2667 ms
User avatar
Angelina Mayo
 
Posts: 3427
Joined: Wed Jan 24, 2007 4:58 am

Post » Sun May 20, 2012 1:57 am

Intel Core i7 930 OC@ 4209.0 MHz

C:\Users\Invisible Black\Desktop\New folder (5)>echo off---------- testing started ----------testing x87Starting matrix multiplication loop...Elapsed time: 1872 mstesting sse1Starting matrix multiplication loop...Elapsed time: 1903 mstesting sse2Starting matrix multiplication loop...Elapsed time: 1763 ms---------- testing finished ----------Press any key to continue . . .
User avatar
Becky Cox
 
Posts: 3389
Joined: Thu Jun 22, 2006 8:38 am

Post » Sat May 19, 2012 7:58 pm

Althon II X3 435 Rana @ 2.91 ghz

---------- testing started ----------testing x87Starting matrix multiplication loop...Elapsed time: 4094 mstestingStarting matrix multiplication loop...Elapsed time: 5610 mstestingStarting matrix multiplication loop...Elapsed time: 5453 ms---------- testing finished ----------Press any key to continue . . .
User avatar
Kay O'Hara
 
Posts: 3366
Joined: Sun Jan 14, 2007 8:04 pm

Post » Sat May 19, 2012 8:59 pm

Intel Core 2 Duo P8600 2.4Ghz

---------- testing started ----------
testing x87

Starting matrix multiplication loop...
Elapsed time: 4524 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 3931 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 3494 ms
---------- testing finished ----------
Press any key to continue . . .
User avatar
Marlo Stanfield
 
Posts: 3432
Joined: Wed May 16, 2007 11:00 pm

Post » Sat May 19, 2012 11:44 pm

Intel Core2Qaud Q6600 2.4Ghz.


---------- testing started ----------
testing x87

Starting matrix multiplication loop...
Elapsed time: 5195 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 4493 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 3728 ms
---------- testing finished ----------
Press any key to continue . . .
User avatar
kitten maciver
 
Posts: 3472
Joined: Fri Jun 30, 2006 2:36 pm

Post » Sun May 20, 2012 1:24 am

AMD Phenom II X4 965 BE OC@ 3.9Ghz


---------- testing started ----------
testing x87

Starting matrix multiplication loop...
Elapsed time: 3073 ms
testing sse1

Starting matrix multiplication loop...
Elapsed time: 4103 ms
testing sse2

Starting matrix multiplication loop...
Elapsed time: 4134 ms
---------- testing finished ----------
Press any key to continue . . .

User avatar
Isabella X
 
Posts: 3373
Joined: Sat Dec 02, 2006 3:44 am

Post » Sun May 20, 2012 8:52 am

Intel Core 2 Quad CPU Q6600 @ 2.40GHz
---------- testing started ----------
testing x87
Starting matrix multiplication loop...
Elapsed time: 5273 ms
testing sse1
Starting matrix multiplication loop...
Elapsed time: 4181 ms
testing sse2
Starting matrix multiplication loop...
Elapsed time: 3588 ms
---------- testing finished ----------

Used Sandra to test CPU heat sensors which report 72 to 74 deg C !!! Time for a clean out or new fan.
User avatar
Rich O'Brien
 
Posts: 3381
Joined: Thu Jun 14, 2007 3:53 am

Post » Sun May 20, 2012 1:44 am

I find it kind of funny how everyone jumped all over them for using x87 code, but that wasn't that bad of a decision to begin with apparently. It appears to be best for AMD machines, and only slightly worse than SSE2 for Intel machines.

theres a difference between x87 and x87 optimisations.

Because there is a bunch of other stuff the mod is doing that doesn't involve math.

inline optimisation is ALL about math.

Something to keep in mind: This is more of a want to see the differences between those three stats per system, not really trying to compare systems themselves to one another. The other is that the more things open or going on in the background the higher the numbers will be. Even turning off firefox with the skyrimnexus page open dropped my numbers about 100ms. Again though: This is for comparising the different methods, not the different systems. It is not an [censored] benchmark.

ED: wait, e-youknowhat is censored? hahahaha.

Nope, background tasks has very little impact on the variation
User avatar
Star Dunkels Macmillan
 
Posts: 3421
Joined: Thu Aug 31, 2006 4:00 pm

Post » Sat May 19, 2012 10:25 pm

-delete
User avatar
kelly thomson
 
Posts: 3380
Joined: Thu Jun 22, 2006 12:18 pm

Post » Sun May 20, 2012 4:27 am

-delete-
User avatar
Beth Belcher
 
Posts: 3393
Joined: Tue Jun 13, 2006 1:39 pm

Post » Sat May 19, 2012 7:27 pm

Continued in http://www.gamesas.com/topic/1331790-skyboost-topic-5/
User avatar
Euan
 
Posts: 3376
Joined: Mon May 14, 2007 3:34 pm

PreviousNext

Return to V - Skyrim