![]() |
||
|
||
|
|
bf3Net TCP/IP Performance FiguresTo give an impression of the system performance of the bf3Net TCP/IP
Protocol Stack, Ethernet datalinksThe chart below displays the data throughputs and (program/data) memory requirements of the bf3Net TCP/IP Protocol Stack for communication through Ethernet datalinks.
(1) WI: Windmill Innovations 100BaseTx Ethernet Daughtercard; EMAC:
DM64x-internal Ethernet MAC;
NWK: bf3Net F2812 Networking Kit.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Target
Board
|
Unit | ||||||
| Parameter | C5402 DSK | 54CST EVM |
C6713 DSK | ||||
| CPU frequency (1) | 100 | 118 | 150 | MHz | |||
| Communications interface (1) | V.90 | CST | V.90 | - | |||
| Max. TCP rx throughput, no-copy | n.a. | - | - | kBytes/s | |||
| Max. TCP rx throughput, copy | 4.9 | - | - | kBytes/s | |||
| Max. TCP tx throughput | 4.9 | - | - | kBytes/s | |||
| Code efficiency - TCP rx, no-copy | n.a. | - | - | CPU cycles/bit | |||
| Code efficiency - TCP rx, copy | 120 | - | - | CPU cycles/bit | |||
| Code efficiency - TCP tx | 120 | - | - | CPU cycles/bit | |||
| Program Memory (min.) | 42.3 | 42.3 | - | kByte (2) | |||
| Program Memory (max.) | 66.6 | 66.6 | - | kByte (2) | |||
| Data Memory (typical) | 11 | 11 | - | kByte (2) | |||
(1) V.90: RS-232 V.90 voiceband modem; CST: 54CST on-chip modem algorithms.
(2) For C5000 platforms (word-addressable CPU) the unit is kWord.
The bf3Net performance figures are obtained as follows:
| Communication Throughput: | The communication throughput figures are measured in MBit/s or kByte/s at TCP level. The following parameters are measured: Max. TCP receive (rx) throughput, no-copy: The mamixum no-copy TCP receive throughput is measured by sending TCP packets with TCP packet generator to a DSP target board running bf3Net, or by downloading a large file from a Microsoft Internet Information Server 5.0 on the local network. Received data is immediately discarded. While this parameter gives an impression of the maximum achievable throughput of the communications interface, it has little or no meaning for any practical purposes. Windmill Innovations provides this number as reference only. Max. TCP receive (rx) throughput, copy: Max. TCP transmit (tx) throughput: Max. UDP transmit (tx) throughput: |
| Code Efficiency Factors: | Across the entire bandwidth, the bf3Net TCP/IP Protocol Stack approximately displays a linear relationship between data throughput and processor load. This relationship is expressed numerically with the so-called code efficiency factors, in "CPU cycles per transferred bit". Using the code efficiency factor, a faithful estimation of the CPU load at a certain TCP bitrate can be made, by multiplying the code efficiency factor with the TCP bitrate. The code efficiency figures are provided for each communication throughput parameter (see above): Code efficiency
- TCP receive (rx), no-copy: Code efficiency - TCP receive (rx), copy: Code efficiency - TCP transmit (tx): Code efficiency - UDP transmit (tx): |
| Memory Requirements: | The bf3Net TCP/IP Protocol Stack is highly modular and configurable. To give an indication of the memory requirements, the following figures are provided: Program memory (min.): Program memory (max.): Data memory (typical): |
Please contact us if you need more detailed results.
Search | Home | Site Map | News | Sales | Technical Support | Contact Info | DSP | Imaging | FPGA | FireWire | Ethernet Traquair reserves the right to change specifications without notice. |