Project

General

Profile

OsmoTRX » History » Version 42

laforge, 02/21/2016 10:44 AM

1 41 sylvain
{{>toc}}
2 1 ttsou
3 41 sylvain
h1. [[OsmoTRX]]
4 1 ttsou
5 1 ttsou
6 41 sylvain
[[OsmoTRX]] is a software-defined radio transceiver that implements the Layer 1 physical layer of a BTS comprising the following 3GPP specifications:
7 41 sylvain
* TS 05.01 "Physical layer on the radio path"
8 41 sylvain
* TS 05.02 "Multiplexing and Multiple Access on the Radio Path"
9 41 sylvain
* TS 05.04 "Modulation"
10 41 sylvain
* TS 05.10 "Radio subsystem synchronization"
11 1 ttsou
12 41 sylvain
[[OsmoTRX]] is based on the transceiver code from the [[OpenBTS]] project, but setup to operate independently with the purpose of using with non-OpenBTS software and projects, while still maintaining backwards compatibility with [[OpenBTS]]. Currently there are numerous features contained in [[OsmoTRX]] that extend the functionality of the [[OpenBTS]] transceiver. These features include enhanced support for various embedded platforms - notably ARM - and dual channel diversity support for the Fairwaves [[UmTRX]].
13 41 sylvain
14 41 sylvain
15 41 sylvain
h2. Features
16 41 sylvain
17 41 sylvain
18 41 sylvain
*Intel SSE Support*
19 6 ttsou
* SSE3
20 6 ttsou
* SSE4.1
21 20 ttsou
22 41 sylvain
On Intel processors, [[OsmoTRX]] makes heavy use of the Streaming SIMD Extensions (SSE) instruction set. Accelerated operations include pulse shape filtering, resampling, sequence correlation, and many other signal processing operations. SSE3 is the minimum requirement for accelerated use.
23 1 ttsou
24 20 ttsou
SSE3 is present in the majority of Intel processors since later versions of the Pentium 4 architecture and is also present on low power Atom processors. Support is automatically detected at build time. For additional performance information, please see the performance and benchmarks section.
25 29 ttsou
26 41 sylvain
*ARM Support*
27 1 ttsou
* NEON
28 1 ttsou
* NEON-VFPv4
29 20 ttsou
30 41 sylvain
[[OsmoTRX]] runs on a variety of ARM processors with and without NEON coprocessors. Like SSE on Intel processors, NEON provides acceleration with SIMD vectorized instructions.
31 20 ttsou
32 1 ttsou
Tested popular architectures include ARM11 (Raspberry Pi), Cortex-A8 (!BeagleBoard), and Cortex-A15 (!ArndaleBoard). Loosely speaking, these platforms are representative of low cost embedded devices, mid-level handsets, and high-end smartphones respectively. Similarly, in order, these platforms include no NEON coprocessor, standard NEON, and NEON-VFPv4. The latter NEON variation, VFPv4, provides additional fused-multiply-accumulate (FMA) instructions useful for many DSP operations.
33 1 ttsou
34 26 ttsou
NEON support must be enabled by the user at build time. For additional information, please see the configuration and performance and benchmarks sections.
35 37 ttsou
36 41 sylvain
*Dual Channel (UmTRX and B210)*
37 7 ttsou
38 1 ttsou
Two dual channel modes are available: standard dual channel mode and diversity. In standard dual channel mode, each RF
39 28 ttsou
path of the dual channel device supports a different ARFCN. Each path operates independently a
40 1 ttsou
nd operates similarly to two separate devices. GSM channel capacity in this mode is doubled. This option can be configured at run time from the command line.
41 1 ttsou
42 41 sylvain
*Dual Channel Diversity (UmTRX, experimental)*
43 1 ttsou
44 28 ttsou
Diversity mode is similar to the standard dual channel mode except each antenna supports both ARFCN channels. In this case, the receiver sample bandwidth is widened to handle both ARFCN's and subsequently converted and demultiplexed into separate sample streams. Each GSM receive path is fed dual signals, where antenna selection diversity is performed by taking the stronger signal on a burst-by-burst basis. This diversity setup improves uplink reception performance in multipath fading environments.
45 16 ttsou
46 28 ttsou
Limitations are increased CPU utilization and that ARFCN spacing is restricted (currently at 400 kHz) by the receiver sampling bandwidth. Setting the ARFCN spacing beyond the sampling limit will disable the diversity path and operate in standard dual channel mode. This options can be configured at run time from the command line.
47 20 ttsou
48 41 sylvain
*Uplink Burst Detection*
49 39 ttsou
50 41 sylvain
[[OsmoTRX]] utilizes an updated receive burst detection algorithm that provides greater sensitivity and reliability than the original [[OpenBTS]] approach, which relied on energy detection for the initial stage of burst acquisition.
51 39 ttsou
52 1 ttsou
The limitation of the previous approach was that it was slow to adapt to highly transient power levels and false burst detection in challenging situations such as receiver saturation, which may occur in close range lab testing. The other issue was that a high degree of level tuning was often necessary to operate reliably.
53 1 ttsou
54 1 ttsou
The current receiver code addressed those limitations for improved performance in a wider variety of environments.
55 1 ttsou
56 41 sylvain
*Low Phase Error Modulator*
57 16 ttsou
58 16 ttsou
The default GSM downlink signal is configured for low distortion using a linearized GMSK modulator. The implementation is based on a two pulse Laurent approximation of continuous phase modulated (CPM) signals. The baseband output signal measures with very low phase error and is capable of passing industry spectrum mask requirements. Please note that actual performance will depend strongly on the particular device in use.
59 1 ttsou
60 41 sylvain
Theoretical details can be found in the report on "GMSK":http://tsou.cc/gsm/report_gmsk.pdf. Octave / Matlab code for "pulse generation":http://tsou.cc/gsm/laurent.m is also available.
61 1 ttsou
62 1 ttsou
This option can be enabled or disabled at run time from the command line.
63 28 ttsou
64 28 ttsou
Very Low Phase Error (Ettus Research N200)
65 16 ttsou
66 42 laforge
!http://tsou.cc/gsm/osmo-trx-phase75.gif!
67 1 ttsou
68 21 ttsou
Spectrum Mask (Ettus Research N200)
69 1 ttsou
70 42 laforge
!http://tsou.cc/gsm/osmo-trx-spectrum75.gif!
71 1 ttsou
72 1 ttsou
73 41 sylvain
h2. RF Hardware support
74 1 ttsou
75 41 sylvain
76 41 sylvain
Multiple RF devices are currently supported. These include USRP family products from Ettus Research, and the [[UmTRX]] from Fairwaves.
77 41 sylvain
78 41 sylvain
||*Fairwaves*||*Notes*||
79 20 ttsou
||UmTRX||Dual channel||
80 20 ttsou
81 20 ttsou
All Ettus Research devices are supported.
82 1 ttsou
83 41 sylvain
||*Ettus Research*||*Notes*||
84 1 ttsou
||USRP1||Requires legacy libusrp driver and clocking modification||
85 1 ttsou
||USRP2||10 MHz external reference required||
86 1 ttsou
||B100||
87 1 ttsou
||B110||
88 1 ttsou
||B200||GPSDO or 10 MHz external reference recommended||
89 1 ttsou
||B210||Dual channel, 10 MHz external reference recommended||
90 1 ttsou
||N200||
91 20 ttsou
||N210||
92 1 ttsou
||E100||
93 1 ttsou
||E110||
94 1 ttsou
95 1 ttsou
96 41 sylvain
h2. Embedded Platform Support
97 1 ttsou
98 41 sylvain
99 41 sylvain
[[OsmoTRX]] has been tested on the multiple embedded platforms representing a wide range of device types. Low cost ARM devices are generally limited by memory and I/O as much CPU utilization.
100 41 sylvain
101 1 ttsou
Running a full or near full ARFCN configuration (7 simultaneous TCH channels with Combination V) may require running the GSM stack remotely, which can be configured at runtime on the command line. This limitation appears to be scheduling related more so than lack of CPU resources, and may be resolved at a later time.
102 20 ttsou
103 41 sylvain
||*Platform*||*SoC*||*Processor*||*SIMD/FPU*||*Testing Notes*
104 20 ttsou
||!ArndaleBoard||Samsung Exynos 5250||ARM Cortex-A15||NEON-VFPv4||7 TCH||
105 41 sylvain
||!BeagleBoard-xM||Texas Instruments OMAP3||ARM Cortex-A8||NEON||7 TCH, remote [[OsmoBTS]] stack||
106 41 sylvain
||Ettus E100||Texas Instruments OMAP3||ARM Cortex-A8||NEON||7 TCH, remote [[OsmoBTS]] stack||
107 41 sylvain
||Raspberry Pi||Broadcom BCM2835||ARM11||VFP||2 TCH, remote [[OsmoBTS]] stack||
108 1 ttsou
||Shuttle PC||NA||Intel Atom D2550||SSE3||Dual channel, 15 TCH||
109 21 ttsou
110 1 ttsou
All embedded plaforms were tested with low-phase error modulator disabled. Use of the more accurate modulator on embedded platforms has not been extensively tested.
111 1 ttsou
112 19 ttsou
113 41 sylvain
h2. Mailing List
114 22 ttsou
115 41 sylvain
116 41 sylvain
For development purposes, [[OsmoTRX]] is discussed on both [[OpenBTS]] and [[OpenBSC]] mailing lists at openbts-discuss@lists.sourceforge.net and openbsc@lists.osmocom.org respectively.
117 41 sylvain
118 1 ttsou
Please direct questions and bug reports to the list appropriate for the GSM stack being used.
119 1 ttsou
120 41 sylvain
Subscription information is available at "and [http://lists.osmocom.org/mailman/listinfo/openbsc/":https://lists.sourceforge.net/lists/listinfo/openbts-discuss].
121 1 ttsou
122 1 ttsou
123 41 sylvain
h2. GPRS support
124 1 ttsou
125 1 ttsou
126 41 sylvain
[[OsmoTRX]] supports GPRS through [[OsmoBTS]].
127 1 ttsou
128 41 sylvain
For GPRS support with [[OpenBTS]], please use the transceiver supplied with [[OpenBTS]].
129 41 sylvain
130 41 sylvain
131 41 sylvain
h2. Source code
132 41 sylvain
133 41 sylvain
134 1 ttsou
The source code is available from git.osmocom.org (module osmo-trx).
135 18 ttsou
136 18 ttsou
Public read-only access is available via
137 41 sylvain
<pre>
138 19 ttsou
$ git clone git://git.osmocom.org/osmo-trx
139 41 sylvain
</pre>
140 1 ttsou
You can browse it via cgit: http://cgit.osmocom.org/cgit/osmo-trx/
141 1 ttsou
142 1 ttsou
143 41 sylvain
h2. Configuration and Build
144 1 ttsou
145 41 sylvain
146 41 sylvain
The only package dependency is the Universal Hardware Driver (UHD), which is available from Ettus Research or Fairwaves depending on the device. Please note that the UHD implementation must match hardware (i.e. Ettus Research UHD for USRP devices and Fairwaves UHD with [[UmTRX]]). The one device that does not use the UHD driver is the USRP1, which is supported through the legacy libusrp driver provided in GNU Radio 3.4.2.
147 41 sylvain
148 36 9600
First, run autoreconf to remake the build system files.
149 41 sylvain
<pre>
150 1 ttsou
$ autoreconf -i
151 18 ttsou
...
152 41 sylvain
</pre>
153 18 ttsou
154 41 sylvain
*Intel Platforms (All)*
155 18 ttsou
156 1 ttsou
Intel SSE support is automatically detected on Intel x86 platforms. No user intervention is necessary. The general configuration defaults to the low phase error modulator. Atom users may wish to use the low-CPU utilization modulator, which can be later enabled from the command line at runtime.
157 41 sylvain
<pre>
158 18 ttsou
$ ./configure
159 1 ttsou
...
160 1 ttsou
checking whether mmx is supported... yes
161 19 ttsou
checking whether sse is supported... yes
162 18 ttsou
checking whether sse2 is supported... yes
163 18 ttsou
checking whether sse3 is supported... yes
164 18 ttsou
checking whether ssse3 is supported... yes
165 18 ttsou
checking whether sse4.1 is supported... yes
166 18 ttsou
checking whether sse4.2 is supported... yes
167 18 ttsou
...
168 41 sylvain
</pre>
169 18 ttsou
170 41 sylvain
*ARM Platforms with NEON*
171 18 ttsou
172 41 sylvain
Many popular ARM development boards fall under this category including BeagleBoard, PandaBoard, and Ettus E100 USRP. This option will disable the low phase error modulator, which can be re-enabled at runtime. NEON support must be manually enabled.
173 41 sylvain
<pre>
174 24 ttsou
$ ./configure --with-neon
175 41 sylvain
</pre>
176 1 ttsou
177 41 sylvain
*ARM Platforms with NEON-VFPv4*
178 1 ttsou
179 41 sylvain
Currently very few development platforms support this instruction set, which is seen mainly in high end smartphones and tablets. Available development boards are ArndaleBoard and ODROID-XU. This option will disable the low phase error modulator, which can be re-enabled at runtime. NEON-VFPv4 support must be manually enabled.
180 41 sylvain
<pre>
181 1 ttsou
$ ./configure --with-neon-vfpv4
182 41 sylvain
</pre>
183 1 ttsou
184 41 sylvain
*ARM Platforms without NEON*
185 1 ttsou
186 41 sylvain
This configuration mainly targets the Raspberry Pi. ARM platforms without NEON vector units are almost always very slow processors, and generally not very suitable for running [[OsmoTRX]]. Running [[OsmoTRX]] on a Raspberry Pi, however, is possible along with limited TCH (voice) channel support. Currently this configuration requires minor code changes.
187 1 ttsou
188 1 ttsou
Coming soon...
189 1 ttsou
190 41 sylvain
*Build and Install*
191 1 ttsou
192 16 ttsou
After configuration, installation is simple.
193 16 ttsou
194 41 sylvain
<pre>
195 16 ttsou
$ make
196 16 ttsou
$ sudo make install
197 41 sylvain
</pre>
198 16 ttsou
199 16 ttsou
200 41 sylvain
h2. Running
201 16 ttsou
202 41 sylvain
203 41 sylvain
[[OsmoTRX]] can be configured with a variety of options on the command line. In most cases, the default settings will suffice. Notable options include UHD device argument passing, which is often useful for using network based devices with firewalls, and external 10 MHz reference support.
204 41 sylvain
205 41 sylvain
<pre>
206 16 ttsou
$ osmo-trx -h
207 16 ttsou
linux; GNU C++ version 4.8.1 20130603 (Red Hat 4.8.1-1); Boost_105300; UHD_003.005.004-140-gfb32ed16
208 16 ttsou
209 16 ttsou
Options:
210 16 ttsou
  -h    This text
211 1 ttsou
  -a    UHD device args
212 16 ttsou
  -l    Logging level (EMERG, ALERT, CRT, ERR, WARNING, NOTICE, INFO, DEBUG)
213 16 ttsou
  -i    IP address of GSM core
214 1 ttsou
  -p    Base port number
215 1 ttsou
  -d    Enable dual channel diversity receiver
216 16 ttsou
  -x    Enable external 10 MHz reference
217 16 ttsou
  -s    Samples-per-symbol (1 or 4)
218 38 ttsou
  -c    Number of ARFCN channels (default=1)
219 38 ttsou
  -f    Enable C0 filler table
220 16 ttsou
  -o    Set baseband frequency offset (default=auto)
221 41 sylvain
</pre>
222 16 ttsou
223 41 sylvain
<pre>
224 1 ttsou
$ osmo-trx -a "addr=192.168.10.2"
225 16 ttsou
linux; GNU C++ version 4.8.1 20130603 (Red Hat 4.8.1-1); Boost_105300; UHD_003.004.000-b14cde5
226 16 ttsou
227 16 ttsou
Config Settings
228 16 ttsou
   Log Level............... INFO
229 1 ttsou
   Device args............. addr=192.168.10.2
230 16 ttsou
   TRX Base Port........... 5700
231 1 ttsou
   TRX Address............. 127.0.0.1
232 16 ttsou
   Channels................ 1
233 16 ttsou
   Samples-per-Symbol...... 4
234 16 ttsou
   External Reference...... Disabled
235 16 ttsou
   Diversity............... Disabled
236 16 ttsou
237 41 sylvain
-- Opening a [[UmTRX]] device...
238 13 ttsou
-- Current recv frame size: 1472 bytes
239 38 ttsou
-- Current send frame size: 1472 bytes
240 41 sylvain
-- Setting [[UmTRX]] 4 SPS
241 38 ttsou
-- Transceiver active with 1 channel(s)
242 41 sylvain
</pre>
243 38 ttsou
244 1 ttsou
245 41 sylvain
h2. [[OsmoTRX]] with [[OpenBTS]]
246 38 ttsou
247 38 ttsou
248 41 sylvain
[[OsmoTRX]] is fully compatible with [[OpenBTS]] for voice and SMS services. Due to differences in handing of GPRS, [[OsmoTRX]] does not support GPRS when used with [[OpenBTS]], however, GPRS with the Osmocom stack is supported.
249 41 sylvain
250 41 sylvain
For use with [[OpenBTS]], enable the filler table option "Enable C0 filler table", which enables [[OpenBTS]] style idle bursts and retransmissions.
251 41 sylvain
252 41 sylvain
<pre>
253 1 ttsou
$ osmo-trx -f
254 41 sylvain
</pre>
255 17 ttsou
256 41 sylvain
The [[OsmoTRX]] transceiver should be started before running [[OpenBTS]]. No symbolic link to './transceiver' should exist in the [[OpenBTS]] directory. This prevents [[OpenBTS]] from starting its own transceiver instance.
257 35 ttsou
258 1 ttsou
259 41 sylvain
h2. Benchmarks
260 1 ttsou
261 35 ttsou
262 41 sylvain
A variety of performance benchmarks are available for various code optimizations. These include floating point - integer conversions, convolution, and convolutional decoding. Note that convolutional decoding does not take place in [[OsmoTRX]], but one stop higher in the Layer 1 stack - either in [[OsmoBTS]] or [[OpenBTS]] core.
263 35 ttsou
264 41 sylvain
*Repository*
265 41 sylvain
266 41 sylvain
Currently the trx-bench repository holds the test files and contains the same NEON and SSE code as [[OsmoTRX]]. The test code may be merged into [[OsmoTRX]] at a later time, but, for now, it exists as a separate repository. NEON configure options are the same as [[OsmoTRX]].
267 41 sylvain
268 41 sylvain
<pre>
269 35 ttsou
$ git clone https://github.com/ttsou/trx-bench.git
270 35 ttsou
271 35 ttsou
$ cd trx-bench
272 35 ttsou
$ autoreconf -i
273 35 ttsou
$ ./configure [--with-neon] [--with-neon-vfp4]
274 1 ttsou
$ make
275 1 ttsou
$ src/conv_test
276 35 ttsou
$ src/convert_test
277 35 ttsou
$ src/convolve_test
278 41 sylvain
</pre>
279 35 ttsou
280 35 ttsou
The convolutional decoding test includes command options including experimental support for benchmarking with multiple threads.
281 35 ttsou
282 41 sylvain
<pre>
283 35 ttsou
$ ./conv_test -h
284 35 ttsou
Options:
285 35 ttsou
  -h    This text
286 1 ttsou
  -i    Number of iterations
287 1 ttsou
  -j    Number of threads for benchmark (1 to 32)
288 13 ttsou
  -b    Run benchmark tests
289 13 ttsou
  -a    Run validity checks
290 13 ttsou
  -e    Run bit-error-rate tests
291 41 sylvain
</pre>
292 10 ttsou
293 1 ttsou
Selected benchmark results are provided below. All tests are run on a single core only.
294 1 ttsou
295 41 sylvain
*Intel Haswell (i7 4770K 3.5 GHz)*
296 1 ttsou
297 41 sylvain
<pre>
298 1 ttsou
--- Floating point to integer conversions
299 1 ttsou
-- Testing 40000 iterations of 3120 values
300 1 ttsou
- Measuring conversion time
301 1 ttsou
- Elapsed time base...                  0.065508 secs
302 1 ttsou
- Validating SIMD conversion results... PASS
303 3 ttsou
- Measuring conversion time
304 3 ttsou
- Elapsed time SIMD ...                 0.011424 secs
305 3 ttsou
- Speedup...                            5.734244
306 41 sylvain
</pre>
307 1 ttsou
308 41 sylvain
<pre>
309 3 ttsou
[+] Testing: GSM TCH/AFS 7.95 (recursive, flushed, punctured)
310 3 ttsou
[.] Input length  : ret = 165  exp = 165 -> OK
311 3 ttsou
[.] Output length : ret = 448  exp = 448 -> OK
312 3 ttsou
[.] Pre computed vector checks:
313 3 ttsou
[..] Encoding: OK
314 3 ttsou
[..] Decoding base: 
315 3 ttsou
[..] Decoding SIMD: 
316 1 ttsou
[..] Code N 3
317 1 ttsou
[..] Code K 7
318 1 ttsou
OK
319 1 ttsou
[.] Random vector checks:
320 1 ttsou
[.] Testing baseline:
321 17 ttsou
[..] Encoding / Decoding 10000 cycles:
322 17 ttsou
[.] Elapsed time........................ 1.435066 secs
323 17 ttsou
[.] Rate................................ 3.121808 Mbps
324 17 ttsou
[.] Testing SIMD:
325 17 ttsou
[..] Encoding / Decoding 10000 cycles:
326 17 ttsou
[.] Elapsed time........................ 0.073524 secs
327 17 ttsou
[.] Rate................................ 60.932485 Mbps
328 17 ttsou
[.] Speedup............................. 19.518334
329 41 sylvain
</pre>
330 17 ttsou
331 41 sylvain
*Intel Atom (D2500 1.86 GHz)*
332 41 sylvain
<pre>
333 17 ttsou
--- Floating point to integer conversions
334 17 ttsou
-- Testing 40000 iterations of 3120 values
335 17 ttsou
- Measuring conversion time
336 17 ttsou
- Elapsed time base...                 1.147449 secs
337 1 ttsou
- Validating SSE conversion results... PASS
338 17 ttsou
- Measuring conversion time
339 1 ttsou
- Elapsed time SSE ...                 0.347838 secs
340 1 ttsou
- Quotient...                          3.298803
341 41 sylvain
</pre>
342 17 ttsou
343 41 sylvain
<pre>
344 1 ttsou
[+] Testing: GSM TCH/AFS 7.95 (recursive, flushed, punctured)
345 1 ttsou
[.] Input length  : ret = 165  exp = 165 -> OK
346 1 ttsou
[.] Output length : ret = 448  exp = 448 -> OK
347 1 ttsou
[.] Pre computed vector checks:
348 17 ttsou
[..] Encoding: OK
349 17 ttsou
[..] Decoding base: 
350 17 ttsou
[..] Decoding SIMD: 
351 17 ttsou
[..] Code N 3
352 17 ttsou
[..] Code K 7
353 17 ttsou
OK
354 17 ttsou
[.] Random vector checks:
355 19 ttsou
[.] Testing baseline:
356 19 ttsou
[..] Encoding / Decoding 10000 cycles:
357 19 ttsou
[.] Elapsed time........................ 11.822688 secs
358 17 ttsou
[.] Rate................................ 0.378932 Mbps
359 17 ttsou
[.] Testing SIMD:
360 17 ttsou
[..] Encoding / Decoding 10000 cycles:
361 17 ttsou
[.] Elapsed time........................ 0.550423 secs
362 17 ttsou
[.] Rate................................ 8.139195 Mbps
363 17 ttsou
[.] Speedup............................. 21.479277
364 41 sylvain
</pre>
365 17 ttsou
366 41 sylvain
*!ArndaleBoard (ARM Cortex-A15 1.7 GHz)*
367 17 ttsou
368 17 ttsou
Please note that the Viterbi implementations on ARM is largely C based with speedup generated primarily through algorithm changes. In comparison, vector optimization on Intel platforms with SSE is currently much more aggressive, which explains the disparity on decoding performance.
369 17 ttsou
370 41 sylvain
<pre>
371 17 ttsou
--- Floating point to integer conversions
372 17 ttsou
-- Testing 40000 iterations of 3120 values
373 17 ttsou
- Measuring conversion time
374 17 ttsou
- Elapsed time base...                 0.384097 secs
375 17 ttsou
- Validating SSE conversion results... PASS
376 17 ttsou
- Measuring conversion time
377 17 ttsou
- Elapsed time SSE ...                 0.100877 secs
378 17 ttsou
- Quotient...                          3.807578
379 41 sylvain
</pre>
380 17 ttsou
381 41 sylvain
<pre>
382 17 ttsou
[+] Testing: GSM TCH/AFS 7.95 (recursive, flushed, punctured)
383 17 ttsou
[.] Input length  : ret = 165  exp = 165 -> OK
384 17 ttsou
[.] Output length : ret = 448  exp = 448 -> OK
385 17 ttsou
[.] Pre computed vector checks:
386 17 ttsou
[..] Encoding: OK
387 17 ttsou
[..] Decoding base: 
388 17 ttsou
[..] Decoding SIMD: 
389 17 ttsou
[..] Code N 3
390 17 ttsou
[..] Code K 7
391 17 ttsou
OK
392 17 ttsou
[.] Random vector checks:
393 17 ttsou
[.] Testing baseline:
394 17 ttsou
[..] Encoding / Decoding 10000 cycles:
395 17 ttsou
[.] Elapsed time........................ 5.371288 secs
396 17 ttsou
[.] Rate................................ 0.834064 Mbps
397 3 ttsou
[.] Testing SIMD:
398 3 ttsou
[..] Encoding / Decoding 10000 cycles:
399 3 ttsou
[.] Elapsed time........................ 1.016621 secs
400 3 ttsou
[.] Rate................................ 4.406755 Mbps
401 3 ttsou
[.] Speedup............................. 5.283471
402 41 sylvain
</pre>
403 3 ttsou
404 41 sylvain
*!BeagleBoard-xM (ARM Cortex-A8 800 MHz)*
405 41 sylvain
<pre>
406 5 ttsou
--- Floating point to integer conversions
407 3 ttsou
-- Testing 40000 iterations of 3120 values
408 3 ttsou
- Measuring conversion time
409 3 ttsou
- Elapsed time base...                  6.292542 secs
410 4 ttsou
- Validating SIMD conversion results... PASS
411 3 ttsou
- Measuring conversion time
412 3 ttsou
- Elapsed time SIMD ...                 0.839081 secs
413 3 ttsou
- Quotient...                           7.499326
414 41 sylvain
</pre>
415 1 ttsou
416 41 sylvain
<pre>
417 31 ttsou
[+] Testing: GSM TCH/AFS 7.95 (recursive, flushed, punctured)
418 31 ttsou
[.] Input length  : ret = 165  exp = 165 -> OK
419 31 ttsou
[.] Output length : ret = 448  exp = 448 -> OK
420 31 ttsou
[.] Pre computed vector checks:
421 31 ttsou
[..] Encoding: OK
422 31 ttsou
[..] Decoding base: 
423 31 ttsou
[..] Decoding SIMD: 
424 31 ttsou
[..] Code N 3
425 31 ttsou
[..] Code K 7
426 1 ttsou
OK
427 32 ttsou
[.] Random vector checks:
428 32 ttsou
[.] Testing baseline:
429 32 ttsou
[..] Encoding / Decoding 10000 cycles:
430 32 ttsou
[.] Elapsed time........................ 21.963257 secs
431 32 ttsou
[.] Rate................................ 0.203977 Mbps
432 32 ttsou
[.] Testing SIMD:
433 32 ttsou
[..] Encoding / Decoding 10000 cycles:
434 32 ttsou
[.] Elapsed time........................ 3.083282 secs
435 32 ttsou
[.] Rate................................ 1.452997 Mbps
436 32 ttsou
[.] Speedup............................. 7.123337
437 41 sylvain
</pre>
438 32 ttsou
439 32 ttsou
440 41 sylvain
*Full Results*
441 32 ttsou
442 41 sylvain
"[http://tsou.cc/gsm/shuttle.txt":http://tsou.cc/gsm/haswell.txt]
443 31 ttsou
444 41 sylvain
"[http://tsou.cc/gsm/beagle.txt":http://tsou.cc/gsm/arndale.txt]
445 1 ttsou
446 30 ttsou
447 1 ttsou
448 41 sylvain
h2. Authors
449 1 ttsou
450 1 ttsou
451 41 sylvain
[[OsmoTRX]] is currently developed and maintained by Thomas Tsou with generous support from Fairwaves, the Open Technology Institute, and Ettus Research. The code is derived from the [[OpenBTS]] project, which was originally developed by David Burgess and Harvind Samra at Range Networks.
Add picture from clipboard (Maximum size: 48.8 MB)