We resume our examination of operation of professional accelerators
in professional programs. This time we will carry out tests in another
CAD system from Unigraphics.
Last time we studied accelerators in the Solid Edge system, and today we
have a similar system named SolidWorks,
which will demonstrate operation of video cards from a little different
standpoint.
Tested cards
In the tests I used all video cards we have in our lab,
as well as one "gaming" card. I must say that Matrox models are not pure
gaming solutions. It seems that the company regards their cards the same
way because you can hardly play comfortably on this card. I guess you understand
that I'm speaking of the Matrox
Parhelia. Why have I included it into the list? Matrox has recently
released new drivers which passed the certification for some CAD systems,
including SolidWorks 2001. This lets me hope that certification for other
professional applications will also appear soon. As the the card and drivers
are certified for professional applications we can use it on an equal footing
with other solutions of this market segment, especially because its price
makes it even closer to its professional siblings.
The following driver versions were used:
The ATI FireGL2 was tested with the ver.2088, the ATI FireGL8x00 -
ver.3036. The NVIDIA based cards were tested with the latest official available
version - 4041. The 3Dlabs Wildcat VP870 ran under the ver. 0423. The VP870
was estimated both in the maximum geometry mode and in the maximum texture
one. The Parhelia was used with the drivers 1.1.0.80.
Test system configuration
The cards were tested on the Intel's platform based on the Intel Pentium
4 clocked at 2533 MHz, ASUS P4T533 motherboard with 32bit RDRAM 512 MB
PC4200, and Seagate Barracuda ATA IV hard drive of 40 GB. The ViewSonic
P810 and ViewSonic P817-E, 21" each, were used as "visualizers".
Testing technique
The testing technique is unsophisticated - the only thing you have to do it to
write down the results and carefully follow the test terms&conditions. A tested
accelerator is inserted into the test system, then and the Windows XP Professional
OS is installed anew, as well as the Service Pack 1. After that we install all
necessary drivers for the hardware components. It's followed by installation of
the automatic design and engineering system SolidWorks 2001 with the Typical set.
The developers didn't provide any benchmark with the system, that is why we used
tests from www.spec.org.
The benchmark can be downloaded from http://www.spec.org/gpc/apc.static/prev_benchmarks.html.
Originally the test was written by the experts from www.solidworks.com
in Visual Basic and C for Windows NT and Windows 2000.
After that it was optimized by programmers from SPECapc.
The test involves most functions frequently used in designer's daily work. The
source codes of the test are open, and you can look through them. If you want
to test your own system make sure that the installed SoldWorks 2001 is patched
with the Service Pack 5 or a higher version. To avoid any problems it's better
begin the test right after start-up of the system. After every test you should
reboot the system.
Like the Solid Edge, the today's test has very strong requirements for
the system, and I recommend that you read the help file first. All the
tests were carried out in 1280x1024 at 32bit color.
The test consists of several subtests dealing with a processor, video
system and so on. We are certainly interested in the video system test.
There are three test stages. To compare the cards we calculated the arithmetic
mean out of three possible values. The results obtained strongly depended
on something we couldn't identify. Nevertheless, I think that the arithmetic
mean does reflect the truth.
In the driver settings I switched off vertical synchronization, line
anti-aliasing and enabled optimization for the SolidWorks where it
was possible.
Tests
First I will shows you screenshots of the test scenes so that you have an idea
what's it's all about. ;-) It's a real problem to get them as according to the
test terms we couldn't even move the mouse. But I still managed to get several
screenshots. They are obtained on different cards as there is no noticeable difference
between them, that is why you may think that they are received with the card you
are mostly interested in.
The output data are quite complicated; here are the results:
sw2001.Result.graphics = 4.19
sw2001.Result.graphics0.BottomCase.Time = 74.66
sw2001.Result.graphics0.BottomCase.Ratio = 4.53
sw2001.Result.graphics1.AssemblyLine.Time = 5.23
sw2001.Result.graphics1.AssemblyLine.Ratio = 4.67
sw2001.Result.graphics2.AssemblyLineTransparent.Time = 1.86
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio = 5.38
sw2001.Result.graphics3.AirplaneEngine.Time = 30.30
sw2001.Result.graphics3.AirplaneEngine.Ratio = 3.66
sw2001.Result.graphics4.SuzukiEngine.Time = 66.06
sw2001.Result.graphics4.SuzukiEngine.Ratio = 5.20
sw2001.Result.graphics5.SuzukiEngineHLR.Time = 53.83
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio = 2.50
Here I chose the first variable "sw2001.Result.graphics". You will understand
why from the readme file for the test. Later you will see complete data
for all cards in consideration with all these values. The table also shows
arithmetic mean values as it makes no sense to include all data.
So, here is the diagram we are going to use to compare the accelerators.
Unfortunately, the comparison is not pure as the test greatly depends on
a central processor, even in spite of the fact that the SPECaps carried
out the video test. But even here we can clearly see that the Matrox's
efforts to make the certification of the drivers and card for professional
applications are futile. The quality is good, but the speed... Although such
a poor result is not a disaster yet, it causes worry because exactly this
driver was first to get certification for operation with the SolidWorks.
The Quadro 2 copes with its task equally to the Quadro 4. None of the
Quadro cards has problems with quality. The FireGL line is in the same
situation, and the FireGL 2 proves again that it's not too old yet.
Although the numbers are not great, they still show that even an owner
of the Quadro 2 MXR can consider himself on owner of the Quadro 4 900 in
operation in the SolidWorks 2001 :)
Finally, here is the table I promised you; I guess the results need
no commenting on.
FireGL 2:
sw2001.Result.graphics0.BottomCase.Time |
76.33 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.43 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.41 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.52 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
2.17 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
4.60 |
sw2001.Result.graphics3.AirplaneEngine.Time |
30.45 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.64 |
sw2001.Result.graphics4.SuzukiEngine.Time |
71.80 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
4.78 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
31.31 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
4.30 |
FireGL 8700:
sw2001.Result.graphics0.BottomCase.Time |
80.02 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.23 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.67 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.31 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
2.51 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
3.99 |
sw2001.Result.graphics3.AirplaneEngine.Time |
30.46 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.65 |
sw2001.Result.graphics4.SuzukiEngine.Time |
70.3 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
4.89 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
32.68 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
4.27 |
FireGL 8800:
sw2001.Result.graphics0.BottomCase.Time |
80.15 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.23 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.70 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.28 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
2.50 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
4.00 |
sw2001.Result.graphics3.AirplaneEngine.Time |
30.31 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.66 |
sw2001.Result.graphics4.SuzukiEngine.Time |
68.92 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
4.98 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
29.37 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
4.67 |
Parhelia:
sw2001.Result.graphics0.BottomCase.Time |
262.96 |
sw2001.Result.graphics0.BottomCase.Ratio |
1.29 |
sw2001.Result.graphics1.AssemblyLine.Time |
9.34 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
2.62 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
3.29 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
3.05 |
sw2001.Result.graphics3.AirplaneEngine.Time |
67.1 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
1.65 |
sw2001.Result.graphics4.SuzukiEngine.Time |
275.58 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
1.27 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
70.83 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
1.92 |
Quadro 2 MXR:
sw2001.Result.graphics0.BottomCase.Time |
75.2 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.50 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.19 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.71 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
2.33 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
4.30 |
sw2001.Result.graphics3.AirplaneEngine.Time |
31.09 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.56 |
sw2001.Result.graphics4.SuzukiEngine.Time |
65.39 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
5.25 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
26.83 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
5.02 |
Quadro 2 PRO:
sw2001.Result.graphics0.BottomCase.Time |
72.91 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.64 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.20 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.69 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
2.22 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
4.51 |
sw2001.Result.graphics3.AirplaneEngine.Time |
31.45 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.52 |
sw2001.Result.graphics4.SuzukiEngine.Time |
60.45 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
5.68 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
26.88 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
5.02 |
Quadro DCC:
sw2001.Result.graphics0.BottomCase.Time |
72.94 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.64 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.17 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.72 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
2.22 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
4.51 |
sw2001.Result.graphics3.AirplaneEngine.Time |
31.48 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.52 |
sw2001.Result.graphics4.SuzukiEngine.Time |
72.03 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
4.77 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
26.48 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
5.09 |
Quadro 4 550:
sw2001.Result.graphics0.BottomCase.Time |
73.03 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.63 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.27 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.64 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
1.95 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
5.12 |
sw2001.Result.graphics3.AirplaneEngine.Time |
32.31 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.43 |
sw2001.Result.graphics4.SuzukiEngine.Time |
59.59 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
5.76 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
29.83 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
4.52 |
Quadro 4 700:
sw2001.Result.graphics0.BottomCase.Time |
73.56 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.60 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.19 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.71 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
2.41 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
4.16 |
sw2001.Result.graphics3.AirplaneEngine.Time |
30.55 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.63 |
sw2001.Result.graphics4.SuzukiEngine.Time |
60.11 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
5.72 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
27.86 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
4.86 |
Quadro 4 750:
sw2001.Result.graphics0.BottomCase.Time |
72.73 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.66 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.16 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.74 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
2.22 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
4.51 |
sw2001.Result.graphics3.AirplaneEngine.Time |
29.91 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.71 |
sw2001.Result.graphics4.SuzukiEngine.Time |
56.71 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
6.06 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
26.26 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
5.14 |
Quadro 4 900:
sw2001.Result.graphics0.BottomCase.Time |
72.51 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.67 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.16 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.73 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
2.30 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
4.35 |
sw2001.Result.graphics3.AirplaneEngine.Time |
29.91 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.71 |
sw2001.Result.graphics4.SuzukiEngine.Time |
57.03 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
6.03 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
27.04 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
5.00 |
VP870 Geometry:
sw2001.Result.graphics0.BottomCase.Time |
75.17 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.51 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.22 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.68 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
1.99 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
5.04 |
sw2001.Result.graphics3.AirplaneEngine.Time |
31.77 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.49 |
sw2001.Result.graphics4.SuzukiEngine.Time |
68.28 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
5.03 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
50.90 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
2.71 |
VP870 Textures:
sw2001.Result.graphics0.BottomCase.Time |
74.89 |
sw2001.Result.graphics0.BottomCase.Ratio |
4.52 |
sw2001.Result.graphics1.AssemblyLine.Time |
5.13 |
sw2001.Result.graphics1.AssemblyLine.Ratio |
4.76 |
sw2001.Result.graphics2.AssemblyLineTransparent.Time |
1.95 |
sw2001.Result.graphics2.AssemblyLineTransparent.Ratio |
5.15 |
sw2001.Result.graphics3.AirplaneEngine.Time |
30.79 |
sw2001.Result.graphics3.AirplaneEngine.Ratio |
3.60 |
sw2001.Result.graphics4.SuzukiEngine.Time |
67.77 |
sw2001.Result.graphics4.SuzukiEngine.Ratio |
5.07 |
sw2001.Result.graphics5.SuzukiEngineHLR.Time |
55.32 |
sw2001.Result.graphics5.SuzukiEngineHLR.Ratio |
2.44 |
So, the test measures task fulfillment time and calculates a certain
score in accordance with the results obtained. After all tasks are completed
the test calculates a general score which is shown in the summary diagram.
The above tables clearly show where the cards perform well and where they
lose.
Write a comment below. No registration needed!