|
MULTICORE FOR BASE STATION MODEMS
|
| bet | 10/14 | | Sana | 23.07.2024 | | Hajmi | 0,76 Mb. | | #268329 |
Bog'liq TrendsInMulticoreDSP.Gatherer MULTICORE FOR BASE STATION MODEMS
Finally, the last five years have seen many multicore entrants into the base station modem business for cellular infrastructure. The most successful have been DSP-based with a modest number of CPUs and significant shared resources in memory, accelera- tion, and I/O. An example of such a device is the TI TCI6487 shown in Figure 11.
Applications that use these multicore devices require very tight latency constraints, and each core often has a unique func- tionality on the chip. For instance, one
core might do only transmit while another does receive and another does symbol rate processing. Again, this is not a generic programming problem. Each core has a specific and very well-timed set of tasks to perform. The trick is to make sure that timing and performance issues do not occur due to the sharing of non- CPU resources [38].
However, the base-station market also attracted new multicore architectures in a way that neither handset (where the cost constraints and volume tended to favor hardwired solutions beyond the ARM/DSP platform) nor transcoding (where the complexity of the software has kept “stan- dard” DSP multicore in the forefront) have experienced. Examples of these new para- digm companies include Chameleon, PACT, BOPS, Picochip, Morpho, Morphics, and Quicksilver. These companies arose
units (ALUs), with a central controller, and arrays of small CPUs, tightly connected and generally intended to communi- cate in a very synchronized manner. Figure 8 shows the picoAr- ray used by picoChip, a proponent of regular, meshed arrays of processors. Serious programming challenges remain with this kind of architecture because it requires two distinct modes of programming, one for the CPUs themselves and one for the interconnect between the CPUs. A single programming lan- guage would have to be able to not only partition the workload but also comprehend the memory locality, which is severe in a mesh-based architecture.
|
| |
