If you like BoxMatrix then please contribute Supportdata, Supportdata2, Firmware and/or Hardware (get in touch).
My metamonk@yahoo.com is not reachable by me since years. Please use hippie2000@webnmail.de instead.
Property:do read pipe (procfs)
| BoxMatrix >> System >> do_read_pipe (procfs) | @ BoxMatrix - IRC-Chat - Translate: de es fr it nl pl |
| News | Selectors | Models | Accessories | Components | Environment | Config | Commands | System | Webif | Software | Develop | Lexicon | Community | Project | Media |
| Devices | Filesystems | Partitions | Sockets | Netlink | Pipes | Interfaces | Bridges | Ports | Events | Sources | Sinks | AVMIPC | Processes | Watchdogs | Memory | Slab | Vmalloc | ProcFS | SysFS | Research |
 |
Name-Collision - multiple objects in this wiki use the name do_read_pipe!
|
Procfs-Object
| Procfs-Object: | do_read_pipe (procfs) - type File | Wiki | Freetz | IPPF | whmf | AVM | Web |
| Location: | System >> Proc-Filesystem - Origin: AVM | ||||||
| Path: | /proc/avm, /proc/avm/benchmark | ||||||
| Properties: | Firmware: 5.04 - 7.58 - Kernel (perm): 2.6.28.10 (r), 2.6.32.21 (r), 2.6.32.60 (r), 2.6.32.61 (r), 2.6.39.4 (r), 3.10.107 (r), 3.10.73 (r), 4.1.38 (r), 4.1.52 (r), 4.4.271 (r), 4.4.60 (r) | ||||||
| Function: | Test: MemBench Linear Read Benchmark. | ||||||
Goto: Examples - Dependencies - Model-Matrix - Help Supportdata2 - SMW-Browser
Details
do_read_pipe is an open source RAM Benchmark by AVM which performs a Linear Read test (Load-Multiple).
Initially located at /proc/avm it was moved to /proc/avm/benchmark later. See the Path in the Model-Matrix.
This test reads from continous memory locations so it is CPU pipeline friendly.
do_read_pipe is part of the MemBench suite. For an overview of all tests see help (procfs) and the Benchmark article.
Excerpts from the test results (german):
6490 ATOM: Lineares Lesen aus dem RAM 6490 ARM: Burstartiges Lesen aus dem RAM unter Nutzung von load multiple. 7240 / 7390 / 7490 MIPS: Lesen aus dem RAM mit optimaler Unterstuetzung der Pipline. D.h. der Code ist gewaehlt, dass die Pipeline nicht geleert werden muss und so keine Zeit verschwendet wird.
Excerpt from the source code avm_membench.c (german):
/*------------------------------------------------------------------------------------------*\ * Pipeline-friendly Read * * -16x 4-byte-Werte pro Schleifendurchlauf * -> 16 Lesezugriffe pro Schleifendurchlauf \*------------------------------------------------------------------------------------------*/
Examples
7530 fw 7.29 example:
root@fritz1:/var/mod/root# cat /proc/avm/benchmark/do_read_pipe cat: read error: Cannot allocate memory
6490 fw 7.29 example - ATOM side:
root@fritz2:/var/mod/root# cat /proc/avm/benchmark/do_read_pipe AVM-RAM-Benchmark ============================================= IRQs: off (alle Tests mit deaktivierten IRQs) CPU-Clock: 1200042000 RAM-Clock: 100000000 (eff. Datentaktrate) BUS-Breite (Word=): 32 Bit Measure-Time: 1 * 1.0s -- Results -- ============================================================================= type ! total read ! loops ! DDR-Ticks ! 32Bit ! ! in kb ! ! /32Bit ! Worte/s ! kB/s ============================================================================= read ! 554496 ! 1 ! 0.704 ! 141950976 ! 554496 ! ! ! ! ! Lineares Lesen aus dem RAM -----------------------------------------------------------------------------
6490 fw 7.29 example - ARM side:
root@fritz2:/var/mod/root# rpc cat /proc/avm/benchmark/do_read_extreme AVM-RAM-Benchmark ============================================= IRQs: off (alle Tests mit deaktivierten IRQs) CPU-Clock: 450000000 RAM-Clock: 250000000 (eff. Datentaktrate) BUS-Breite (Word=): 32 Bit Measure-Time: 1 * 1.0s -- Results -- ============================================================================= type ! total read ! loops ! DDR-Ticks ! 32Bit ! ! in kb ! ! /32Bit ! Worte/s ! kB/s ============================================================================= read ! 122304 ! 1 ! 7.984 ! 31309824 ! 122304 ! ! ! ! ! Burstartiges Lesen aus dem RAM unter Nutzung von load multiple. -----------------------------------------------------------------------------
7490 fw 7.29 example:
fritz3:/var/mod/root $ cat /proc/avm/do_read_pipe AVM-RAM-Benchmark ============================================= IRQs: off (alle Tests mit deaktivierten IRQs) CPU-Clock: 500000000 RAM-Clock: 500000000 (eff. Datentaktrate) BUS-Breite (Word=): 16 Bit Measure-Time: 1 * 1.0s -- Results -- ======================================================================================== type ! total read ! loops ! DDR-Ticks ! 16Bit ! ! Stalls per ! ! in kb ! ! /16Bit ! Worte/s ! kB/s ! Instruction ! ======================================================================================== read ! 264192 ! 1 ! 3.696 ! 135266304 ! 264192 ! 7.958 ! Pipeline-friendly ! ! ! ! ! ! ! Lesen aus dem RAM mit optimaler Unterstuetzung der Pipline. ! ! ! D.h. der Code ist gewaehlt, dass die Pipeline nicht geleert ! ! ! werden muss und so keine Zeit verschwendet wird. ! ! ! ----------------------------------------------------------------------------------------
7390 fw 6.04 example:
root@7390:/var/mod/root# cat /proc/avm/do_read_pipe AVM-RAM-Benchmark ============================================= IRQs: off (alle Tests mit deaktivierten IRQs) CPU-Clock: 500000000 RAM-Clock: 333333332 (eff. Datentaktrate) BUS-Breite (Word=): 32 Bit Measure-Time: 1 * 1.0s -- Results -- ============================================================================= type ! total read ! loops ! DDR-Ticks ! 32Bit ! ! in kb ! ! /32Bit ! Worte/s ! kB/s ============================================================================= read ! 145472 ! 1 ! 8.950 ! 37240832 ! 145472 Pipeline-friendly ! ! ! ! ! Lesen aus dem RAM mit optimaler Unterstuetzung der Pipline. ! D.h. der Code ist gewaehlt, dass die Pipeline nicht geleert ! werden muss und so keine Zeit verschwendet wird. ! -----------------------------------------------------------------------------
7240 fw 6.06 example:
root@fritz:/var/mod/root# cat /proc/avm/do_read_pipe AVM-RAM-Benchmark ============================================= IRQs: off (alle Tests mit deaktivierten IRQs) CPU-Clock: 360000000 RAM-Clock: 120000000 (eff. Datentaktrate) BUS-Breite (Word=): 16 Bit Measure-Time: 1 * 1.0s -- Results -- ============================================================================= type ! total read ! loops ! DDR-Ticks ! 16Bit ! ! in kb ! ! /16Bit ! Worte/s ! kB/s ============================================================================= read ! 133056 ! 1 ! 1.761 ! 68124672 ! 133056 Pipeline-friendly ! ! ! ! ! Lesen aus dem RAM mit optimaler Unterstuetzung der Pipline. ! D.h. der Code ist gewaehlt, dass die Pipeline nicht geleert ! werden muss und so keine Zeit verschwendet wird. ! -----------------------------------------------------------------------------
Dependencies
Daily updated index of all dependencies of this object. Last update: GMT.
A ** in the Mod column marks info from Supportdata2 probes, which will always stay incomplete.
A - in the Mod column marks manual research, the Firmware then shows where the Object occurs, not the Relation.
| Relation | Typ | Object | Mod | Firmware | Info | Origin |
|---|---|---|---|---|---|---|
| 0 dependencies for this object | ||||||
Model-Matrix
Daily updated index of the presence, path and size of this object for each model. Last update: 2024-11-13 05:19 GMT.
Showing all models using this object. Click any column header (click-wait-click) to sort the list by the respective data.
The (main/scrpn/boot/arm/prx/atom/rtl) label in the Model column shows which CPU is meant for Multi-Linux models.
Note that this list comes from Supportdata2 probes, which can have arbitrary settings and come from different firmware versions.
It doesn't say much if a model is not listed here. It may be a missing supportdata2 file or just a disabled feature.
| Model | Firmware | Path | Kernel (perm) |
|---|---|---|---|
| FRITZ!Box 3490 | 7.30 | /proc/avm | 3.10.107 (r) |
| FRITZ!Box 4060 | 7.30 - 7.39 | /proc/avm/benchmark | 4.4.271 (r), 4.4.60 (r) |
| FRITZ!Box 5490 | 7.29 | /proc/avm | 3.10.107 (r) |
| FRITZ!Box 5590 Fiber (main) | 7.29 | /proc/avm/benchmark | 4.4.60 (r) |
| FRITZ!Box 6430 Cable (arm) | 7.29 | /proc/avm/benchmark | 2.6.39.4 (r) |
| FRITZ!Box 6430 Cable (atom) | 7.29 | /proc/avm/benchmark | 2.6.39.4 (r) |
| FRITZ!Box 6490 Cable (arm) | 7.29 - 7.39 | /proc/avm/benchmark | 2.6.39.4 (r) |
| FRITZ!Box 6490 Cable (atom) | 7.29 - 7.39 | /proc/avm/benchmark | 2.6.39.4 (r) |
| FRITZ!Box 6850 LTE | 7.39 | /proc/avm/benchmark | 4.4.60 (r) |
| FRITZ!Box 6850 5G | 7.39 | /proc/avm/benchmark | 4.4.60 (r) |
| FRITZ!Box Fon WLAN 7240 | 6.06 | /proc/avm | 2.6.32.60 (r) |
| FRITZ!Box 7272 | 6.88 | /proc/avm/benchmark | 2.6.32.61 (r) |
| FRITZ!Box Fon WLAN 7320 | 6.55 | /proc/avm/benchmark | 2.6.32.61 (r) |
| FRITZ!Box 7362 SL | 7.12 | /proc/avm | 3.10.107 (r) |
| FRITZ!Box Fon WLAN 7390 | 6.04 | /proc/avm | 2.6.28.10 (r) |
| FRITZ!Box 7412 | 6.87 | /proc/avm | 3.10.73 (r) |
| FRITZ!Box 7430 | 7.29 | /proc/avm | 3.10.107 (r) |
| FRITZ!Box 7490 | 7.29 - 7.51 | /proc/avm | 3.10.107 (r) |
| FRITZ!Box 7510 | 7.30 | /proc/avm/benchmark | 4.4.60 (r) |
| FRITZ!Box 7520 | 7.29 - 7.50 | /proc/avm/benchmark | 4.4.271 (r), 4.4.60 (r) |
| FRITZ!Box 7520 v2 | 7.31 | /proc/avm/benchmark | 4.4.60 (r) |
| FRITZ!Box 7530 | 7.39 | /proc/avm/benchmark | 4.4.271 (r) |
| FRITZ!Box 7530 AX | 7.31 | /proc/avm/benchmark | 4.1.52 (r) |
| FRITZ!Box 7581 | 7.17 | /proc/avm/benchmark | 4.1.38 (r) |
| FRITZ!Smart Gateway | 7.57 - 7.58 | /proc/avm/benchmark | 4.4.271 (r) |
| FRITZ!Repeater 1200 AX | 7.30 - 7.39 | /proc/avm/benchmark | 4.4.60 (r) |
| FRITZ!Repeater 6000 | 7.29 - 7.39 | /proc/avm/benchmark | 4.4.60 (r) |
| Speedport W 504V MK | 5.04 | /proc/avm | 2.6.32.21 (r) |
| 28 models use this object | |||
Help Supportdata2
The data in this article is incomplete since it was manually collected using the Supportdata2 project.
Unlike the Supportdata-Probes which have been collected for years Supportdata2 is brand new and only has a few probes.
If you have access to a shell then please help to extend the Supportdata2 collection to improve this data.
It's easy and it's done in minutes. Please send created data as an Email attachment to the address listed here. Thanks!
