maths coprocessor This is a topic that many people are looking for. bluevelvetrestaurant.com is a channel providing useful information about learning, life, digital marketing and online courses …. it will help you have an overview and solid multi-faceted knowledge . Today, bluevelvetrestaurant.com would like to introduce to you 80387 Math Coprocessors Speed Comparison. Following along are instructions in the video below:
“And welcome to another episode from us here digital collection about history in this episode. Episode. I want to talk about mathematical processors. They were all the rage in the world from around mid 80s to mid nineties.
The advent of the intel 4 8. 6. Dx processor. Which included math proper functionality inside the processor itself pretty much killed the market for the coprocessors before anything prior to that such as if you have a computer.
We have an intel 8086 286 of 386 and you wanted to run kind of applications for example you needed to purchase the mathematica processors. Intel produced those for each generation receive use to allow applications to perform fast floating point operations. Others. Are the processors at the time.
Say. The one produced by dec corporation. Such as like pdp. 11.
4x. And so on you they also supported floating point processors. So you know there was there was a need for it in the market. What i wanna do today is i want to look at the the later america processors those produced for the 80386 processors and just test the speed say.
Which one is fast. I know if you if you know back you let s say. The early 90s like 90 or 91. What which one would you choose so for today.
I have four of them i have iit 80 c 3 c 87 processor i might have a silex spanish math math coprocessor i also have two versions of the intel s own 83 seven math ones and will run to suite of applications. And then what was the conclusion study and i m missing some most notably the you lsi math coprocessor as well as the intel rapid cam. Now the reason i did include rapid kr is because it s not as much a math of a crystal itself. But rather it.
Updates your processor from a three eight six two four eight six and also adding math property. So that s kind of a separate category so anyway let s let s let s see our contenders you now that we established what microprocessors we are testing. I want explain a little bit about testing methodology. I try to choose applications from a large spectrum of options.
So that they will get the better feeling the words faster. So first. I chose a application called vista pro. Which is kind of hard like application at the time it allowed you to render somewhat realistic landscapes then i chose synthetic mathematical tasks three of them link back lll which is little more labs loops and whetstone finally.
I also chose have chosen quite for game. Rendering. So will the london post source to all these texan applications. We will see at the end.
What results we have all right so let s talk about our results as i mentioned before i will i wanted it with something like a card application. So i chose vista pro and i did two renderings rate. One of the maui island and one of the matterhorn peak. The first one was six hundred forty nine thousand triangles and second two one three hundred thirty thousand triangles then i chose three synthetic math effects first is lean pack second one is lll which stands for livermore labs loops.
Which runs a lot of math kernels actually. It s the longest. One of the three last is the whetstone benchmark and i wanted to finish with a game application for that i chose quake while not playable it still allowed me to differentiate between that for mathematical processor speed so holy measure is well the first two are measured in second so lower value is better the next three return million floating pointing floating point operations per second or m flops. That live higher is better and last.
One is quake. Returns frames per second and again higher is better alright. So let s look at the first contender. Which was the iit 3c 87.
Coprocessor that render the maui island in 416 seconds. And it rendered the matter core in. 267 seconds for a. Synthetic test a.
Linpack. Return 041 megaflops lll return. 047 and finally west stone report is 057 quake run at an abysmal but doable 15 frames per second then i tested the cyrix fast math chip. This one was 20 faster 400 seconds for maui and 266 seconds for matterhorn.
In linpack i did a lot better. 047 ll was a lot better 062. And surprise. What stone actually scored as you know quantity to megaflops and even in.
Quake as you d expect it is better 18. So that meant that now silex is our fastest chip that we have at the moment then i took a look at the first of the two intel based r87 dx chips the l30 one one four four eight that was slightly slower in maui 417 rendered matterhorn 268 slower. However for synthetic test is called better across the board than the iit chip sometimes better sometimes slower than fast map the quake. Results was also in between the two the last one was the second intel 376 chip.
I had the first one is actually my note as a note of it isis territory megahertz. The second one did not have a one thing to keep in mind is that all those around 40 megahertz. So the slightly slower on on maui. But otherwise pretty much identical results with the previous trade 7dx.
So after all this this how how do we look so how which one is the fastest. So i want to put a caveat here that this is only for the applications of tested. There are other things you can run to test the mathematical speed. The results might be slama sometimes different so in the last place suprisingly enough was actually the iit 3c 87.
Followed by in the third and second place by the intel. The winner of my test turned out to be the cyrix fast math. ” ..
Thank you for watching all the articles on the topic 80387 Math Coprocessors Speed Comparison. All shares of bluevelvetrestaurant.com are very good. We hope you are satisfied with the article. For any questions, please leave a comment below. Hopefully you guys support our website even more.