Is it possible to add a new algorithm to cryptopp benchmarks?

[Mehmet TUFEKCI]

Hi everyone! I want to bench an algorithm (aes version) which is not in cryptopp benchmarks. If it is possible, how can I do this? Thanks in advance.

[Jeffrey Walton]

On Tuesday, April 11, 2017 at 4:17:06 PM UTC-4, Mehmet TUFEKCI wrote:

Hi everyone! I want to bench an algorithm (aes version) which is not in cryptopp benchmarks. If it is possible, how can I do this? Thanks in advance.

Yes. An example of add a new cipher is available at https://github.com/weidai11/cryptopp/commit/8ca0f479391c54ee454e9cddfcbe1657da240f85.

Ensure the name you use in regtest.cpp (to register the name) is the same name used when bench marking (bench marks create objects based on the registered name).

Jeff

[Mehmet TUFEKCI]

Hi everyone!

I add my new algorithm to crypto++. But when I make bench it gives segmentation fault. Any idea why?

Thanks in advance.

On Mon, Apr 17, 2017 at 9:32 AM, Mehmet TUFEKCI <mtufek...@gmail.com> wrote:

Dear Jeffrey Walton;

Which version of crypto++ is used for this?

Best regards..

Mehmet

--
--
You received this message because you are subscribed to the "Crypto++ Users" Google Group.
To unsubscribe, send an email to cryptopp-users-unsubscribe@googlegroups.com.
More information about Crypto++ and this group is available at http://www.cryptopp.com.
---
You received this message because you are subscribed to a topic in the Google Groups "Crypto++ Users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/cryptopp-users/rwRSMrxJVcw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to cryptopp-users+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jeffrey Walton]

On Monday, April 17, 2017 at 4:59:25 PM UTC-4, Mehmet TUFEKCI wrote:

Hi everyone!

I add my new algorithm to crypto++. But when I make bench it gives segmentation fault. Any idea why?

No idea. Run it under Valgrind or another memory checker.

Jeff

[Mehmet TUFEKCI]

18 Nisan 2017 Salı 09:10:22 UTC+3 tarihinde Jeffrey Walton yazdı:

Hi Jeff;

I have registered my new algorithm. I add aes.cpp to crypto++ library.  But there is problem;

void AES_X::Base::ProcessAndXorBlock(const byte *inBlock, const byte *xorBlock, byte *outBlock) const
{
    xorbuf(outBlock, inBlock, m_x1, BLOCKSIZE);
    //m_aes->ProcessAndXorBlock(outBlock, xorBlock, outBlock);
    xorbuf(outBlock, m_x3, BLOCKSIZE);
}

In this function if I use "m_aes->ProcessAndXorBlock(outBlock, xorBlock, outBlock);" it gives segmentation fault. Any idea why?

Best regards...

[Jeffrey Walton]

> I have registered my new algorithm. I add aes.cpp to crypto++ library. But
> there is problem;
>
> void AES_X::Base::ProcessAndXorBlock(const byte *inBlock, const byte
> *xorBlock, byte *outBlock) const
> {
> xorbuf(outBlock, inBlock, m_x1, BLOCKSIZE);
> //m_aes->ProcessAndXorBlock(outBlock, xorBlock, outBlock);
> xorbuf(outBlock, m_x3, BLOCKSIZE);
> }
>
> In this function if I use "m_aes->ProcessAndXorBlock(outBlock, xorBlock,
> outBlock);" it gives segmentation fault. Any idea why?

No, I have no idea.

Perhaps you can provide the Valgrind trace? Or maybe the stack
associated with the crash?

I don't know anyone who can answer questions like you ask with no
information. https://www.cryptopp.com/wiki/Bug_Report.

Jeff

[Mehmet TUFEKCI]

5 Mayıs 2017 Cuma 00:14:49 UTC+3 tarihinde Jeffrey Walton yazdı:

Jeff;

Here is my valgrind trace added log_test.txt. Could you pls have a look?

Thanks in advance.

[Jeffrey Walton]

On Thu, May 4, 2017 at 5:35 PM, Mehmet TUFEKCI <mtufek...@gmail.com> wrote:
>
> 5 Mayıs 2017 Cuma 00:14:49 UTC+3 tarihinde Jeffrey Walton yazdı:
>>
>> > I have registered my new algorithm. I add aes.cpp to crypto++ library.
>> > But
>> > there is problem;
>> >
>> > void AES_X::Base::ProcessAndXorBlock(const byte *inBlock, const byte
>> > *xorBlock, byte *outBlock) const
>> > {
>> > xorbuf(outBlock, inBlock, m_x1, BLOCKSIZE);
>> > //m_aes->ProcessAndXorBlock(outBlock, xorBlock, outBlock);
>> > xorbuf(outBlock, m_x3, BLOCKSIZE);
>> > }
>> >
>> > In this function if I use "m_aes->ProcessAndXorBlock(outBlock, xorBlock,
>> > outBlock);" it gives segmentation fault. Any idea why?
>>
>> No, I have no idea.
>>
>> Perhaps you can provide the Valgrind trace? Or maybe the stack
>> associated with the crash?
>>
>> I don't know anyone who can answer questions like you ask with no
>> information. https://www.cryptopp.com/wiki/Bug_Report.
>>
>

> Here is my valgrind trace added log_test.txt. Could you pls have a look?

It looks like you have not built with symbols or you are building with
optimizations that are too high. Please visit
http://valgrind.org/docs/manual/quick-start.html#quick-start.prepare.

This usually means you need to update Valgrind:

==6901== by 0x8069413: BenchmarkAll(double, double) (bench1.cpp:302)
==6901== by 0x804DAAC: main (test.cpp:367)
==6901== Your program just tried to execute an instruction that Valgrind
==6901== did not recognise. There are two possible reasons for this.
==6901== 1. Your program has a bug and erroneously jumped to a non-code
==6901== location. If you are running Memcheck and you just saw a
==6901== warning about a bad jump, it's probably your program's fault.
==6901== 2. The instruction is legitimate but Valgrind doesn't handle it,
==6901== i.e. it's Valgrind's fault. If you think this is the case or
==6901== you are not sure, please let us know and we'll try to fix it.
==6901== Either way, Valgrind will now raise a SIGILL signal which will
==6901== probably kill your program.



Jeff

[Jeffrey Walton]

On Thursday, May 4, 2017 at 6:27:56 PM UTC-4, Jeffrey Walton wrote:

On Thu, May 4, 2017 at 5:35 PM, Mehmet TUFEKCI <mtufek...@gmail.com> wrote:
>
> 5 Mayıs 2017 Cuma 00:14:49 UTC+3 tarihinde Jeffrey Walton yazdı:

>> ...

>
> Here is my valgrind trace added log_test.txt. Could you pls have a look?

It looks like you have not built with symbols or you are building with
optimizations that are too high. Please visit
http://valgrind.org/docs/manual/quick-start.html#quick-start.prepare.

This usually means you need to update Valgrind:

==6901==    by 0x8069413: BenchmarkAll(double, double) (bench1.cpp:302)
==6901==    by 0x804DAAC: main (test.cpp:367)
==6901== Your program just tried to execute an instruction that Valgrind
==6901== did not recognise.  There are two possible reasons for this.
==6901== 1. Your program has a bug and erroneously jumped to a non-code
==6901==    location.  If you are running Memcheck and you just saw a
==6901==    warning about a bad jump, it's probably your program's fault.
==6901== 2. The instruction is legitimate but Valgrind doesn't handle it,
==6901==    i.e. it's Valgrind's fault.  If you think this is the case or
==6901==    you are not sure, please let us know and we'll try to fix it.
==6901== Either way, Valgrind will now raise a SIGILL signal which will
==6901== probably kill your program.

We added a recipe to build the library for Valgrind; see https://github.com/weidai11/cryptopp/commit/c62e1ade6ed1e0f6d739dbe6a59f08975c88a60b.

You will still need to update Valgrind or build Valgrind from the project's SVN sources to get the latest VEX decoder. Also see http://valgrind.org/downloads/repository.html.

Nearly every machine I have runs Valgrind from SVN sources. I can't count how many times I've experienced the SIGILL because VEX was out of date from a distro.

Jeff

[Mehmet TUFEKCI]

7 Mayıs 2017 Pazar 05:59:51 UTC+3 tarihinde Jeffrey Walton yazdı:

Dear Jeff;

Thanks for Valgrind update. I have got latest VEX decoder. But nothing changed. I have attaxhed my logs again. On the other hand I have changed my new aes version. I have written aes_x. In aes_x I used rijndael codes but in the aes_x.cpp it's given " ‘‘Se’ was not declared in this scope " error. I have attached my aes_x.cpp and aes_x.h files. Could you please check it and tell where is my problem?

Best regards..
 

[Jeffrey Walton]

Thanks for Valgrind update. I have got latest VEX decoder. But nothing changed. I have attaxhed my logs again. On the other hand I have changed my new aes version. I have written aes_x. In aes_x I used rijndael codes but in the aes_x.cpp it's given " ‘‘Se’ was not declared in this scope " error. I have attached my aes_x.cpp and aes_x.h files. Could you please check it and tell where is my problem

https://stackoverflow.com/questions/23307110/valgrind-gives-an-error-for-nearly-everything-warning-client-switching-stacks

It sounds like (to me) you have a memory error somewhere in your program.

Jeff

[Jeffrey Walton]

Dear Jeff;

Thanks for Valgrind update. I have got latest VEX decoder. But nothing changed.

This looks like your Valgrind issue: https://bugs.kde.org/show_bug.cgi?id=295974.

Use another tool.

Jeff