TCdr8(x64) v TCdr7.56a Personal File-copy Speed Test Results

[sucata]

===========
TEST SYSTEM
===========
Intel Quad Core i7
CPU 920 @ 2.67 GHz
1024KB L2 Cache
133MHz
NEC USB3 PCI card
HDD0: 450GB 10K RPM (Boot)
HDD1: 1.5TB 7.2K RPM (Data)
USB3s: 1.5TB 7.2K RPM (Data)
Test File: 12GB ISO

=======
RESULTS
=======

HDD1-to-HDD0
-------------------
Transfer Rate
+ TCdr7: 100MB/s
+ TCdr8: 100MB/s
CPU Utilization
+ TCdr7: 5%-8%
+ TCdr8: 5%-8%
Transfer Time
+ TCdr7: 2 min
+ TCdr8: 2 min
Result: no difference

USB3-to-HDD0
-------------------
Transfer Rate
+ TCdr7
- Burst: 118MB/s
- Sustained: 100MB/s
+ TCdr8
- Burst: 285MB/s (1 sec.)
- Sustained: 112MB/s
CPU Utilization
+ TCdr7: 4%-14%
+ TCdr8: 4%-12%
Transfer Time
+ TCdr7: 2 min
+ TCdr8: 1.5 min
Result: v8 25% faster

USB3-to-USB3
------------------
Note: 2 HDDs
Transfer Rate
+ TCdr7: 100MB/s
+ TCdr8: 100MB/s
CPU Utilization
+ TCdr7: 2%-12%
+ TCdr8: 4%-12%
Transfer Time
+ TCdr7: 2.25 min
+ TCdr8: 2 min
Result: no difference

==========
CONCLUSION
==========

Uninstall v8 beta
Revert back to (stable) 7.57*

*Never copy large amounts
of data to boot drive so no
v8 advantage.

==========
FOOTNOTE
==========

Also saw no significant
difference when copying
30,000 files.

Test Date:
20120304Sun1225 (U.S.A. E.S.T.)

[Horst.Epp]

1. I don't understand your conclusion if I look on your results ?
2. TC 8 is not designed to be faster than TC 7, its main advantage is the x64 version, among many other improvements.

[sucata]

1. Regarding, "its main advantage is the x64 version":

The "advantage" of x64 for any s/w is the fullest possible utilization (within h/w device/cable limits) of "higher" (above 2GB) memory and CPU multi-core usage. Very few programs are optimized to do this; and, TCdr8 is no exception -- x64, or not.

Acknowledging that a hardware device (HDD) places inherent limits on memory & CPU capacity performance; for example, CDex (open source app) extracts audio from CDs at an absolutely BLISTERING pace utilizing 98% of the memory and CPU 8-cores' capacity.

2. RE "among many other improvements":

For early adopters of TCdr, it now (V7.57) has a fairly stable features base; therefore, these results were shared for those who are MOST interested in the software's ability to copy massive files back and forth between fixed, USB, and portable HDDs in the LEAST amount of time -- no practical difference between x86 & x64 on the test PC.

Footnote: the test PC had 8GB of internal memory.

[Sir_SiLvA]

2sucata
are you seriously one of the "dumb sheeps" that belives the
myth 64Bit = faster?

[umbra]

The "advantage" of x64 for any s/w is the fullest possible utilization (within h/w device/cable limits) of "higher" (above 2GB) memory and CPU multi-core usage.

First, what does AMD64 have to do with multi-core usage? And second, why would TC need so much RAM?

... for example, CDex (open source app) extracts audio from CDs at an absolutely BLISTERING pace utilizing 98% of the memory and CPU 8-cores' capacity.

That would be because its not just copying data from a CD but also (and mainly) transcoding it. That's why it needs so much computer power.

... the software's ability to copy massive files back and forth between fixed, USB, and portable HDDs in the LEAST amount of time ...

That's true, but no software can be faster than the physical drive it's copying from/to.

Footnote: the test PC had 8GB of internal memory.

Irrelevant. A lot of memory is good only for caching, which is already done by Windows. Unless you copy the same file over and over again, it's main advantage is that, in some scenarios, it can limit a speed impact caused by simultaneous file operations. However it won't speed up anything, since you are limited by a performance of your disk (or its interface).

There are some operations, for which TC could use multiple cores - creating checksums, verification of archives, ... But there is no way you would see it in a simple "one file copy" operation.

And to the results from your post - are you sure you haven't made a mistake? There are too many "100MB/s" values. On the other hand, those results seem to be realistic, since in all your tests you were copying to/from a 1+TB disk. The peak speed of these kinds of drives is usually 110-140MB/s (and lower if you connect them through USB3), so I don't think TC could be much faster. Do you have any results showing, that TC is slower than something else?

[sucata]

@"Sir" SiLvA:

"Whoever one is, and wherever one is, one is always in the wrong if one is rude." ~Maurice Baring

Thanks for reminding me why I stopped visiting this forum many years ago. Signing off (for good)...

PS: @ubmra, thanks for taking the time for your thoughtful response.