Benchmark Analysis (Cont.)

SiSoftware Sandra

Configuration 1

• Model: Intel(R) Pentium(R) 4 CPU 2.80GHz
• Speed: 2.80GHz
• Performance Rating: PR3724 (estimated)
• FSB Speed: 2x 200MHz (400MHz data rate)

Sandra Benchmark	Results
Int Buff'diSSE2	4305MB/s
Float Buff'd iSSE2	4316MB/s
Maximum Bus Bandwidth	6400MB/s

Int Buff'd iSSE2 (Integer Stream) Results

Sandra Benchmark	Breakdown
Assignment	4265MB/s
Scaling	4256MB/s
Addition	4355MB/s
Triad	4345MB/s
Data Item Size	16-bytes
Buffering Used	Yes
Offset Displacement Used	Yes
Bandwidth Efficiency	67% (estimated)

Float Buff'd iSSE2 (Float Stream) Results

Sandra Benchmark	Breakdown
Assignment	4270MB/s
Scaling	4277MB/s
Addition	4360MB/s
Triad	4359MB/s
Data Item Size	16-bytes
Buffering Used	Yes
Offset Displacement Used	Yes
Bandwidth Efficiency	67% (estimated)

SiSoftware Sandra

Configuration 2

• Model: Intel(R) Pentium(R) 4 CPU 2.80GHz
• Speed: 3.26GHz
• Performance Rating: PR4338 (estimated)
• FSB Speed: 2x 233MHz (466MHz data rate)

Sandra Benchmark	Results
Int Buff'diSSE2	5333MB/s
Float Buff'd iSSE2	5308MB/s
Maximum Bus Bandwidth	7456MB/s

Int Buff'd iSSE2 (Integer Stream) Results

Sandra Benchmark	Breakdown
Assignment	5232MB/s
Scaling	5231MB/s
Addition	5434MB/s
Triad	5435MB/s
Data Item Size	16-bytes
Buffering Used	Yes
Offset Displacement Used	Yes
Bandwidth Efficiency	72% (estimated)

Float Buff'd iSSE2 (Float Stream) Results

Sandra Benchmark	Breakdown
Assignment	5220MB/s
Scaling	5191MB/s
Addition	5411MB/s
Triad	5411MB/s
Data Item Size	16-bytes
Buffering Used	Yes
Offset Displacement Used	Yes
Bandwidth Efficiency	71% (estimated)

The Sandra benches were ran the real world way, with the test rig being a daily user in our office (review test rig) and LAN gaming rig commonly used on the weekends. Being said the memory benchmarks ran under full Microsoft services and all 28 processes loaded into memory at the time the of the analysis.

This ram gets more efficient, the higher the overclock. On our system the 800MHz FSB was getting about a 67% efficiency rating. After overclocking to 14 x 233MHz the efficiency rating climbed to 72% which is indicative of super low latency.