SpeedSys score of 526.00
L2 = 4881 MB/s
RAM = 217 MB/s
Cyrix DX4-100 CPU
M919 Motherboard, 256 KB Cache, 128 MB FPM RAM
Don't ask me how I did it. A short in the L2 cache? Out of fear of blowing up the cpu/motherboard, I did not attempt a Windows98SE boot.
M919 motherboard? It's PC-CHIPS, anything is possible 😜 🤣
VERY funky memory bandwidth graph there, looks like it didn't pick up the cpu's L1 cache at all?
If you are squeamish, don't prod the beach rubble.
You should try running Quake on that thing 🤣
Its floating-point performance would probably still suck.
I dont see the CPU clock anywhere? At what clocks are you running it? It can't be the 34Mhz, right? 😜
Some sort of buggy timer that things seem to appear quicker than they really are?
Chipset?
1+1=10
wrote:Some sort of buggy timer that things seem to appear quicker than they really are?
Must be one of those curious Cyrix bugs acting up again 😜
And memory bandwidth 100.16 mega bytes per second??
wrote:wrote:Some sort of buggy timer that things seem to appear quicker than they really are?
Must be one of those curious Cyrix bugs acting up again 😜
And memory bandwidth 100.16 mega bytes per second??
Yeah, it all looks really screwed - nearly 5Gbytes of cache bandwidth - and 180Mbytes/sec from an IDE disk. Methinks not 😀
My collection database and technical wiki:
https://www.target-earth.net
Look feipoa, we know you're really into the Cyrix thang and that it means a lot to you, just don't let it get the better of you, alright? ;D 😜
Not sure what Tetrium is going on about -- I am merely sharing an anomaly that occurred on its own, without intervention. The Cyrix DX4 was never a stand-up performer, unlike the Cyrix 5x86's. I do not plan on replicating these results.
To answer some questions,
This Cyrix DX4 runs at 100 Mhz. Speedsys does not always calculate the correct cpu speed, which coincides with Harekiet's comment about a buggered up hardware timer, although I've never seen SpeedSys bugger up on the rating score like this.
L1 cache is present, but it is likely out of the scale's range, probably in the hundreds regime. This leads me to think that something odd occured with the L2 cache alone. I am in general agreement with the DonutKing in that "anything is possible" with PC Chips products.
For the harddrive's linear speed read, the graph is correct, but the printed average value is quite high at 183,661 KB/s. While this may be possible on an Ultra320 SCSI harddrive (the harddrive that was used), the employed SCSI Ultra2 controller is only capable of a maximum 80 MB/s.
The memory bandwidth is consistent with other 486-class cpus's tested (100 Mbytes/sec for 33 Mhz FSB and 121 Mbytes/second for 40 Mhz FSB).
The chipset is UMC 881/886, found in many late era, PCI-based socket 3 motherboards.
Has anyone else seen such elevated speed ratings in SpeedSys?
wrote:Not sure what Tetrium is going on about
Nah, I was just kidding mate 😉
