|
|
SOCKET 7 CPU BENCHMARKS AND
OBSERVATIONS
Last updated: 2/1/2000
Below, we present benchmarks for some popular
Socket 7 CPU's; namely, the AMD 450 Mhz K6-3, 450 Mhz K6-2, 400 Mhz K6-2,
and the VIA Cyrix 300 Mhz M2-433GP. Based on Winstone 99
testing by NSTL (National Software Testing Laboratories Inc.), VIA
rates the 290 Mhz M2-400GP chip's general performance to be a little faster
than a 400 Mhz Celeron and a little slower than a 400 Mhz K6-2. They
don't compare the 300 Mhz M2-433GP chip to a K6-2 because there is no 433
Mhz K6-2. I did. The M2 is little faster in some respects than
a 400 Mhz K6-2 and quite a bit slower in others. The Business Winstone
99 performance of the two chips is almost identical; but, like all Cyrix
chips I've seen, it has a noticeably weak Floating Point Unit. So,
if you are the average business user, you shouldn't see any noticeable difference
between the 400 Mhz K6-2 and the 300 Mhz M2-433GP. If you play games,
run a software music synthesizer, or do other tasks involving intensive number-crunching,
the K6-2 will beat the M2 coming and going.
Some observations...
-
The M2 runs noticeably cooler than any
of the K6 processors (see below). After all, it is clocked slower.
-
The K6-3 beats all of the other processors
hands down. Unfortunately, the K6-3 is being discontinued (or being renamed/revamped
to become the K6-2+). As I understand it, AMD wants to
emphasize the K7 or Athlon and the K6-3 did not catch-on as competitor
with the Intel P3. (I'll bet 'cha there is a Socket 370 CPU in
the works somewhere; especially since VIA is about to release a socket
370 CPU, code named Joshua.)
-
If you scoot down towards the bottom of
this page you'll see that the WinTune 98 Floating Point benchmark on
the 450 Mhz K6-2 beats the 450 Mhz K6-3??? The test was run three
times.
-
My testing shows that all of these chips,
when used with Epox Super7 motherboards, make excellent Windows NT 4.0
workgroup file servers.
The prices I've seen on the M2's and
K6-2's (currently a $14 difference between the most expensive, the M2-433GP,
and the least expensive, the 400 Mhz K6-2) definitely make the 450 Mhz K6-2
a better buy (performance/cost) than either the 400 Mhz K6-2 or M2-433GP. I
plan to benchmark a 500 Mhz K6-2 before too long (after the price drops to
where the 450 is now).
Larry
BENCHMARK
|
AMD
450
Mhz
K6-3 |
AMD
450
Mhz
K6-2 |
AMD
400
Mhz
K6-2 |
VIA
Cyrix
300 Mhz
M2 433GP |
Winstone
99 |
| Business
Winstone 99 (Winstone units) |
20.1 |
16.8 |
16.6 |
16.5 |
WinBench
99 |
| Business
Disk WinMark 99 (Thousand Bytes/Sec) |
2760 |
2450 |
2430 |
2550 |
| Business
Graphics WinMark 99 |
98.5 |
81.9 |
79.7 |
83.3 |
| CPUmark
99 |
49.3 |
25.9 |
25.4 |
26.4 |
| Disk
Playback/Bus:Overall (Thousand Bytes/Sec) |
2760 |
2450 |
2430 |
2550 |
| Disk
Playback/HE:AVS/Express 3.4 (Thousand Bytes/Sec) |
4330 |
4610 |
4550 |
4280 |
| Disk
Playback/HE:FrontPage 98 (Thousand Bytes/Sec) |
40000 |
32700 |
32300 |
33500 |
| Disk
Playback/HE:MicroStation SE (Thousand Bytes/Sec) |
8640 |
8190 |
8250 |
8200 |
| Disk
Playback/HE:Overall (Thousand Bytes/Sec) |
7560 |
7410 |
7410 |
7270 |
| Disk
Playback/HE:Photoshop 4.0 (Thousand Bytes/Sec) |
5750 |
5700 |
5700 |
5680 |
| Disk
Playback/HE:Premiere 4.2 (Thousand Bytes/Sec) |
5690 |
5500 |
5570 |
5610 |
| Disk
Playback/HE:Sound Forge 4.0 (Thousand Bytes/Sec) |
9410 |
9100 |
9150 |
8940 |
| Disk
Playback/HE:Visual C++ 5.0 (Thousand Bytes/Sec) |
10200 |
9290 |
9260 |
8990 |
| FPU
WinMark |
1500 |
1470 |
1310 |
658 |
| GDI
Playback/HE/AVS/Express 3.4 |
30.6 |
27.3 |
26.2 |
28.2 |
| GDI
Playback/HE/FrontPage 98 |
73.7 |
60.8 |
59.3 |
60.3 |
| GDI
Playback/HE/MicroStation SE |
7.78 |
6.23 |
5.76 |
6.19 |
| GDI
Playback/HE/Photoshop 4.0 |
44.1 |
34.5 |
33.8 |
35.9 |
| GDI
Playback/HE/Premiere 4.2 |
55.5 |
39.9 |
39 |
43.1 |
| GDI
Playback/HE/Sound Forge 4.0 |
108 |
82.7 |
79.8 |
86.5 |
| GDI
Playback/HE/Visual C++ 5.0 |
189 |
149 |
145 |
157 |
| High-End
Disk WinMark 99 (Thousand Bytes/Sec) |
7560 |
7410 |
7410 |
7270 |
| High-End
Graphics WinMark 99 |
304 |
244 |
231 |
248 |
WinTune
98 |
| CPU
Integer (MIPS) |
1211.759 |
1001.571 |
886.7306 |
696.541 |
| CPU
Floating Point (MFLOPS) |
505.3782 |
523.7625 |
467.2161 |
239.9767 |
| Video(2D)
(MPixels/s) |
38.24159 |
38.82422 |
38.27995 |
39.43199 |
| Direct3D
(MPixels/s) |
38.98649 |
38.32246 |
38.14111 |
39.73344 |
| OpenGL
(MPixels/s) |
6.423152 |
5.837588 |
5.680404 |
5.975155 |
| Memory
(MB/s) |
786.6126 |
634.5484 |
578.4255 |
422.3156 |
| Cached
Disk (MB/s) |
95.39346 |
66.22693 |
63.66254 |
70.39934 |
| Uncached
Disk (MB/s) |
3.238554 |
3.182848 |
3.14552 |
3.142974 |
CPU
TEMPERATURES |
| CPU
(°C/°F) |
34/93 |
34/93 |
34/93 |
31/87 |
| System
(°C/°F) |
24/75 |
25/77 |
25/77 |
26/78 |
| Difference
(°C/°F) |
10/18 |
9/17 |
9/17 |
5/9 |
WinBench 99 v. 1.1; WinStone 99 v. 1.2
Test Computer:
Processors as shown, Epox MVP3G motherboard, 1 MB cache', 64 MB, 8 ns PC100
memory, Maxtor 9084U2 8.4 GByte, 5,400 RPM ATA/66/33 hard disk, Windows
98 with FAT32 file system (defragged), Liteon 44X CD-ROM, AOpen T9750 AGP
Display Adapter with 4 MB.
NSTL BUSINESS WINSTONE 99 TEST
RESULTS


NSTL test
computer configurations
WinStone and WinBench are registered trademarks
or trademarks of Ziff-Davis Inc. in the U.S. and other countries.
Please see our Contact page
if you have any comments or corrections that would make this article better. Please use our Forums if you need help with a computer or network problem. |
|