Forums

MIPS vs MHz (MCPS)

Started by siva...@gmail.com June 11, 2008
Hi All,

I have a query on representation of performance of DSP algorithms and its complexity. As far as I know, it has always been in terms of MIPS. Considering the advancements in the hardware, the architectures, the pipeline depths, the clocking speeds would it be appropriate to represent the same in terms of MHz (MCPS).

The reason for phrasing this question is that, I notice quite a few people representing the complexity in MCPS especially for the C6x devices. I have not been able to understand the analogy behind this. Could someone explain and throw some light on it?

Thanks & Regards,
Siva
Siva--
I will just mention one reason for preference of MCPS over MIPS,off late.
There may be more and Iam sure others will "chip"-in

DSP software companies have long used performance of their algorithm on a
specific hardware as an USP to sell their wares. When it comes to the
question of which rating to use, marketing divisions who used MIPS figures
for legacy reasons(CISC machines), switched over to MCPS over MIPS ever
since deeply pipelined processors like VLIW came into the market to show
more accurately the load of an algorithm on processor A over processor B.
It is supposedly more "objective" and "fair" to compare, say a MP3 offering
on C64x Vs SHARC when MCPS is used rather than MIPS.

Does that make sense? :-)

--Bhooshan

due to its supposedly objectivity

2]

On Wed, Jun 11, 2008 at 11:48 AM, wrote:

> Hi All,
>
> I have a query on representation of performance of DSP algorithms and its
> complexity. As far as I know, it has always been in terms of MIPS.
> Considering the advancements in the hardware, the architectures, the
> pipeline depths, the clocking speeds would it be appropriate to represent
> the same in terms of MHz (MCPS).
>
> The reason for phrasing this question is that, I notice quite a few people
> representing the complexity in MCPS especially for the C6x devices. I have
> not been able to understand the analogy behind this. Could someone explain
> and throw some light on it?
>
> Thanks & Regards,
> Siva
>
>

--
-----------------------
"I've missed more than 9000 shots in my career.
I've lost almost 300 games. 26 times I've been trusted to take the game
winning shot and missed.
I've failed over and over again in my life.
And that is why I succeed."
-- Michael Jordan
-----------------------
Also given an algorithmic performance expressed in MCPS on processorA and
processorB, MIPS can be easily derived from MCPS but not the other way
around. So it seems a more "descriptive" language.

--Bhooshan

On Wed, Jun 11, 2008 at 4:46 PM, Bhooshan Iyer
wrote:

> Siva--
> I will just mention one reason for preference of MCPS over MIPS,off late.
> There may be more and Iam sure others will "chip"-in
>
> DSP software companies have long used performance of their algorithm on a
> specific hardware as an USP to sell their wares. When it comes to the
> question of which rating to use, marketing divisions who used MIPS figures
> for legacy reasons(CISC machines), switched over to MCPS over MIPS ever
> since deeply pipelined processors like VLIW came into the market to show
> more accurately the load of an algorithm on processor A over processor B.
> It is supposedly more "objective" and "fair" to compare, say a MP3 offering
> on C64x Vs SHARC when MCPS is used rather than MIPS.
>
> Does that make sense? :-)
>
> --Bhooshan
> due to its supposedly objectivity
>
> 2]
>
> On Wed, Jun 11, 2008 at 11:48 AM, wrote:
>
>> Hi All,
>>
>> I have a query on representation of performance of DSP algorithms and its
>> complexity. As far as I know, it has always been in terms of MIPS.
>> Considering the advancements in the hardware, the architectures, the
>> pipeline depths, the clocking speeds would it be appropriate to represent
>> the same in terms of MHz (MCPS).
>>
>> The reason for phrasing this question is that, I notice quite a few people
>> representing the complexity in MCPS especially for the C6x devices. I have
>> not been able to understand the analogy behind this. Could someone explain
>> and throw some light on it?
>>
>> Thanks & Regards,
>> Siva
>>
>> --
> -----------------------
> "I've missed more than 9000 shots in my career.
> I've lost almost 300 games. 26 times I've been trusted to take the game
> winning shot and missed.
> I've failed over and over again in my life.
> And that is why I succeed."
> -- Michael Jordan
> -----------------------

--
-----------------------
"I've missed more than 9000 shots in my career.
I've lost almost 300 games. 26 times I've been trusted to take the game
winning shot and missed.
I've failed over and over again in my life.
And that is why I succeed."
-- Michael Jordan
-----------------------
Hi Bhooshan,

Thanks so much for your reply. Your explanation does drive the point. Thanks once again.

Regards,
Siva

Hi All,
>
>I have a query on representation of performance of DSP algorithms and its complexity. As far as I know, it has always been in terms of MIPS. Considering the advancements in the hardware, the architectures, the pipeline depths, the clocking speeds would it be appropriate to represent the same in terms of MHz (MCPS).
>
>The reason for phrasing this question is that, I notice quite a few people representing the complexity in MCPS especially for the C6x devices. I have not been able to understand the analogy behind this. Could someone explain and throw some light on it?
>
>Thanks & Regards,
>Siva
>