Test 1- Idle Temperatures

All of these programs do a great job of lowering the Idle temperature of the CPU.  The temperatures shown in the first chart are the result of letting the system sit idle for about 20 minutes. The three cooling program's temperatures had stabilized well before the 20 minute time limit had expired, while using no program, the temperature continued to slowly rise, ultimately to 101.1° F. 

 The difference in cooler program temperatures are mostly due to the differences in room temperature, not the ultimate "cooling" ability of each program. Though there differences in the way the programs enable the HLT command, the results of enabling the command are pretty much the same.

My computer room temperature is very hard to regulate exactly.   With winter in full force here, I have rigged an auxiliary heater and thermostat to try to approximate summer temperatures.  Even with a couple of fans to circulate the air, some stratification occurs and the temperature at the computer (floor level) constantly changes by a couple of degrees with air movement.  It is best to judge the increase or decrease in temperature for each individual test, rather than looking for the highest or lowest temp in a column. 

For someone who has had to deal with the searing heat of the AMD K6-233 overclocked, the Celeron 300A is a very nice change.  This chip responds well to a good sized heat sink and good case ventilation, which is reflected in the numbers below. 

The addition of one of the cooler programs brings the temperature down even lower.   Four and a half degrees above room temp for the processor is pretty darn good.   Note that in my system, the difference between running a program and not running it is over 11°F.  Test 2 - WinBench99
In earlier WinBench tests, I stated that the cooler programs did little to degrade performance.  For my system, some of the programs actually helped performance.  At first I was a bit skeptical of the WinBench results, so I ran the tests another three times each.  I received similar results.  The scores reflected here are the average of 3 tests each on 3 seperate installations of Win95 OSR2.  Win98 performance is similar.

Test 3
I have really grown to like this benchmark.  In fact, I tried to contact the company to upgrade from the shareware version to the registered version.  Unfortunately the company seems to have dropped off the face of the web.  If anyone has a working link to U software, please let me know.  The non-working URL I have for them is: http://www.usoftware.com/

UPDATE: Thanks to Jim McLeod, I have a new link to the program.  Try this German site: http://www1.idg.se/mikrodatorn/sharew/95diag.htm  The email for the author of the program is: underwoe@Colorado.Edu

Here again I received some results that were better when running the programs than without.  I was again skeptical of the results and ended up running each program about 10 times over the course of a few days.  The results varied by a point or two, but the pattern remained constant.

Test 4
Early in the testing, I was looking for applications that I used on a daily basis that I could use as benchmarks.  While running CpuIdle 2.4 and transferring a large file from a CD-ROM to one of my hard disks, I noticed that the CPU temperature shot up from around 80° to over 110° in a short period of time.  I hadn't noticed that much heat being generated with the other programs, so I ended up copying the Windows95 folder from CD to hard disk as a test.  File size = 79 MB.

The hard disk was defragged and the CPU was allowed to cool between each test.  Each test was run 3 times and the results were averaged.

The Cd to disk and disk to disk tests for the Celeron were much the same as the earlier tests performed on the AMD.  The "control panel" version of CpuIdle which still uses the VxD was not as effective in keeping the processor from heating up during the transfer of files.

Test 6
I thought I might as well round out this series by checking hard disk to floppy disk transfer temperatures.  File size was 1.38 MB.

Memory
I had some problems with recording memory usage with a utility called Wintop.  It was suggested that I use Norton Utilities System Information - memory function.  Using Norton, I documented the memory usage for the cooler programs.  This utility, unlike Wintop, did not seem to mind what other programs were running in the background or how many times I rebooted.  The memory usage remained constant.  Thank gawd for small favors. :-)  Please note that these are my memory results and that your non-shared and total module memory will vary depending on what other programs you are running and what options you select with the cooler programs.

Memory Used   Chart 3