Intel Software Network - feed

Re: Problems related to Intel Architecture Code Analyzer

Wed, 19 Aug 2009 23:15:56 -0700

Hello everyone!

I have used the IACA for nearly a month and it is great!
It help me to solve many problems that affects the performance of my code...
(and it is simple to use as long as you have the source code in hand)

However, I have noticed something strange...
Even though the analyzer has correctly identified the code for rcpss/rsqrtss,
they are treated as the same as the divss/sqrtss respectively.(As I know,
both of the approximation instructions should run faster than the accurate
ones. However, the analyzer shows they are blocking the divider port for 14
cycles) Is it a bug or something else?

Thank you for your attention!

Hi,

Thanks for the input on Intel(R) Architecture Code Analyzer.

You are correct, the rcp and rsqrt instructions were misclassified as divider operations.
We are working on providing a fix for this issue.

Tal

Re: Problems related to Intel Architecture Code Analyzer

Mon, 24 Aug 2009 12:08:53 -0700

A fixed version is now available. (ver 1.0.2)

Tal

Re: Problems related to Intel Architecture Code Analyzer

Wed, 02 Sep 2009 01:27:37 -0700

[url=http://www.dvdcollects.com/products/Desperate-Housewives-Season-1-5-DVD-Boxset-DVDS-1698.html]Desperate Housewives Season 1-5 DVD Boxset[/url]
[url=http://www.dvdcollects.com/products/24-Hours-Seasons-1-7-DVD-Boxset-DVDS-1684.html]24 Hours Seasons 1-7 DVD Boxset[/url]
[url=http://www.dvdcollects.com/products/Ally-McBeal-Seasons-1-5-DVD-Boxset-DVDS-1264.html]Ally McBeal Seasons 1-5 DVD Boxset[/url]
[url=http://www.dvdcollects.com/products/The-King-of-Queens-Seasons-1-9-DVD-Boxset-DVDS-1744.html]The King of Queens Seasons 1-9 DVD Boxset[/url]
[url=http://www.dvdcollects.com/products/Xena-Warrior-Princess-Season-1-6-DVD-Boxset-DVDS-1418.html]Xena: Warrior Princess Season 1-6 DVD Boxset[/url]

[url=http://www.dvdcollects.com/products/30-Rock-Seasons-1-3-DVD-Boxset-DVDS-1199.html]30 Rock Seasons 1-3 DVD Boxset[/url]
[url=http://www.dvdcollects.com/products/Angel-Seasons-1-5-DVD-Boxset-DVDS-1265.html] Angel Seasons 1-5 DVD Boxset[/url]
[url=http://www.dvdcollects.com/products/Buffy-The-Vampire-Slayer-Seasons-1-7-DVD-Boxset-DVDS-1769.html]Buffy The Vampire Slayer Seasons 1-7 DVD Boxset[/url]
[url=http://www.dvdcollects.com/products/Mistresses-Seasons-1-2-DVD-Boxset-DVDS-1768.html]Mistresses Seasons 1-2 DVD Boxset [/url]
[url=http://www.dvdcollects.com/products/Samantha-Who-Seasons-1-2-DVD-Boxset-DVDS-1762.html]Samantha Who Seasons 1-2 DVD Boxset[/url]

Re: Problems related to Intel Architecture Code Analyzer

Wed, 02 Sep 2009 01:27:43 -0700

Re: Problems related to Intel Architecture Code Analyzer

Mon, 26 Oct 2009 06:46:18 -0700

Hello, I'm trying out the IACA 1.1 for Win32.
Is the following result a bug?

Intel(R) Architecture Code Analyzer Version - 1.1.0
Analyzed File - regrename.obj
Binary Format - 32Bit
Architecture  - Intel(R) AVX

Analysis Report
---------------
Total Throughput: 1 Cycles;             Throughput Bottleneck: FrontEnd, Port0
Total number of Uops bound to ports:  1
Data Dependency Latency:    1 Cycles;   Performance Latency:    1 Cycles

Port Binding in cycles:
-------------------------------------------------------
|  Port  |  0 - DV |  1 |  2 -  D |  3 -  D |  4 |  5 |
-------------------------------------------------------
| Cycles |  1 |  0 |  0 |  0 |  0 |  0 |  0 |  0 |  0 |
-------------------------------------------------------

N  - port number, DV - Divider pipe (on port 0), D - Data fetch pipe (on ports 2 and 3) 
CP - on a critical Data Dependency Path
N  - number of cycles port was bound
X  - other ports that can be used by this instructions
F  - Macro Fusion with the previous instruction occurred
^  - Micro Fusion happened
*  - instruction micro-ops not bound to a port
@  - Intel(R) AVX to Intel(R) SSE code switch, dozens of cycles penalty is expected
!  - instruction not supported, was not accounted in Analysis

| Num of |          Ports pressure in cycles          |    |
|  Uops  |  0 - DV |  1 |  2 -  D |  3 -  D |  4 |  5 |    |
------------------------------------------------------------
|   1    |  1 |    |  X |    |    |    |    |    |  X | CP | pxor xmm0, xmm1
|   0*   |    |    |    |    |    |    |    |    |    |    | pxor xmm2, xmm2
|   0*   |  X |    |  X |    |    |    |    |    |  X | CP | vpxor xmm3, xmm3, xmm4
|   0*   |    |    |    |    |    |    |    |    |    |    | vpxor xmm5, xmm5, xmm5

About "vpxor xmm3, xmm3, xmm4" instruction, it seems must to be decoded to 1 uop.
So I think that the interpretation of the first operand and the third operand is contrary on the IACA.
Other some operations that destination register value absolutely becomes zero-all (ex: xorps, xorpd, psub*, pcmpgt*) are similar too.

In addition, IACA will be segfault by the following code (only 0 uop instruction(s) between markers).

;START_MARKER
mov ebx, 111
db 0x64, 0x67, 0x90

vpxor xmm0, xmm0, xmm0 ;0 uop

;END_MARKER
mov ebx, 222
db 0x64, 0x67, 0x90

Re: Problems related to Intel Architecture Code Analyzer

Mon, 26 Oct 2009 08:32:41 -0700

Quoting - seizh

Hello, I'm trying out the IACA 1.1 for Win32.
Is the following result a bug?
Intel(R) Architecture Code Analyzer Version - 1.1.0
Analyzed File - regrename.obj
Binary Format - 32Bit
Architecture - Intel(R) AVX

Analysis Report
---------------
Total Throughput: 1 Cycles; Throughput Bottleneck: FrontEnd, Port0
Total number of Uops bound to ports: 1
Data Dependency Latency: 1 Cycles; Performance Latency: 1 Cycles

Port Binding in cycles:
-------------------------------------------------------
| Port | 0 - DV | 1 | 2 - D | 3 - D | 4 | 5 |
-------------------------------------------------------
| Cycles | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
-------------------------------------------------------

N - port number, DV - Divider pipe (on port 0), D - Data fetch pipe (on ports 2 and 3)
CP - on a critical Data Dependency Path
N - number of cycles port was bound
X - other ports that can be used by this instructions
F - Macro Fusion with the previous instruction occurred
^ - Micro Fusion happened
* - instruction micro-ops not bound to a port
@ - Intel(R) AVX to Intel(R) SSE code switch, dozens of cycles penalty is expected
! - instruction not supported, was not accounted in Analysis

| Num of | Ports pressure in cycles | |
| Uops | 0 - DV | 1 | 2 - D | 3 - D | 4 | 5 | |
------------------------------------------------------------
| 1 | 1 | | X | | | | | | X | CP | pxor xmm0, xmm1
| 0* | | | | | | | | | | | pxor xmm2, xmm2
| 0* | X | | X | | | | | | X | CP | vpxor xmm3, xmm3, xmm4
| 0* | | | | | | | | | | | vpxor xmm5, xmm5, xmm5
About "vpxor xmm3, xmm3, xmm4" instruction, it seems must to be decoded to 1 uop.
So I think that the interpretation of the first operand and the third operand is contrary on the IACA.
Other some operations that destination register value absolutely becomes zero-all (ex: xorps, xorpd, psub*, pcmpgt*) are similar too.

In addition, IACA will be segfault by the following code (only 0 uop instruction(s) between markers).
;START_MARKER
mov ebx, 111
db 0x64, 0x67, 0x90

vpxor xmm0, xmm0, xmm0 ;0 uop

;END_MARKER
mov ebx, 222
db 0x64, 0x67, 0x90

Hi,

Thank you for your input. The checking of the idiom was incorrect for the AVX version.

I'm currently working on fixing this matter and other matters as well. I will update when a fix will be available on whatif.

Tal

Re: Problems related to Intel Architecture Code Analyzer

Fri, 30 Oct 2009 08:59:02 -0700

Quoting - Tal Uliel (Intel)

thank you i will publish all...

Re: Problems related to Intel Architecture Code Analyzer

Thu, 05 Nov 2009 11:04:14 -0800

Quoting - Tal Uliel (Intel)

A fixed version (1.1.1) was released today.

Tal