StreamIt

Summary of benchmarks and generated code as of July, 2002

All C code for a Pentium IV processor was run on a Intel® Xeon™ 2.20 GHz, 512KB L2 cache, 2 GB RAM. All timing information was gathered with the time command. Here is the contents of the /proc/cpuinfo file.

Benchmark Description lines of code # of constructs in the program Number of filters in the expanded graph

filters pipelines splitjoins feedbackloops

FIR 64 tap FIR 125 5 1 0 0 132

Radar Radar array front-end 549 8 3 6 0 52

Radio FM Radio with an equalizer 525 14 6 4 0 26

Sort 32 element Bitonic Sort 419 4 5 6 0 242

FFT 64 element FFT 200 3 3 2 0 24

Filterbank 8 channel Filterbank 650 9 3 1 1 51

GSM GSM Decoder 2261 26 11 7 2 46

Vocoder 28 channel Vocoder 1964 55 8 12 1 101

3GPP 3GPP Radio Access Protocol 1087 16 10 18 0 48

Benchmark 250 MHz RAW processor C on a 2.2 GHz Intel Pentium IV

StreamIt on 16 tiles C on a single tile

Utilization # of tiles used MFLOPS Throughput (per 10⁵ cycles) Throughput (per 10⁵ cycles) Throughput (per 10⁵ cycles)

FIR 86% 14 815 1188.1 293.5 445.6

Radar 79% 16 1,231 0.52 app. too large 0.041

Radio 73% 16 421 53.9 8.85 14.1

Sort 64% 16 N/A 2,664.4 225.6 239.4

FFT 42% 16 182 2,141.9 468.9 448.5

Filterbank 41% 16 644 256.4 8.9 7.0

GSM 23% 16 N/A 80.9 app. too large 7.76

Vocoder 23% 15 118 8.74 app. too large 3.35

3GPP 18% 16 44 119.6 17.3 65.7

Benchmark	Description	lines of code	# of constructs in the program				Number of filters in the expanded graph
Benchmark	Description	lines of code	filters	pipelines	splitjoins	feedbackloops	Number of filters in the expanded graph
FIR	64 tap FIR	125	5	1	0	0	132
Radar	Radar array front-end	549	8	3	6	0	52
Radio	FM Radio with an equalizer	525	14	6	4	0	26
Sort	32 element Bitonic Sort	419	4	5	6	0	242
FFT	64 element FFT	200	3	3	2	0	24
Filterbank	8 channel Filterbank	650	9	3	1	1	51
GSM	GSM Decoder	2261	26	11	7	2	46
Vocoder	28 channel Vocoder	1964	55	8	12	1	101
3GPP	3GPP Radio Access Protocol	1087	16	10	18	0	48

Benchmark	250 MHz RAW processor					C on a 2.2 GHz Intel Pentium IV
	StreamIt on 16 tiles				C on a single tile	C on a 2.2 GHz Intel Pentium IV
	Utilization	# of tiles used	MFLOPS	Throughput (per 10⁵ cycles)	Throughput (per 10⁵ cycles)	Throughput (per 10⁵ cycles)
FIR	86%	14	815	1188.1	293.5	445.6
Radar	79%	16	1,231	0.52	app. too large	0.041
Radio	73%	16	421	53.9	8.85	14.1
Sort	64%	16	N/A	2,664.4	225.6	239.4
FFT	42%	16	182	2,141.9	468.9	448.5
Filterbank	41%	16	644	256.4	8.9	7.0
GSM	23%	16	N/A	80.9	app. too large	7.76
Vocoder	23%	15	118	8.74	app. too large	3.35
3GPP	18%	16	44	119.6	17.3	65.7