RISC-V optimization and -mtune

I’ve been getting into RISC-V optimization recently. Partly because I got my SiFive VisionFive 2, and partly because unlike x86 the number of RISC-V instructions is so managable that I may actually have a chance at beating the compiler.

I’m optimizing the inner loops of GNURadio, or in other words the volk library. I’ve been getting up to a about a doubling of the speed compared to the compiled C code, depending on the function.

But it got me thinking how far I could tweak the compiler and its options, too.

Yes, I should have done this much sooner.

Many years ago now I built some data processing thing in C++, and thought it ran too slowly. Sure, I did a debug build, but how much slower could that be? Half speed? Nope. 20x slower.

Of course this time I never compared to a debug build, so don’t expect that kind of difference. Don’t expect that it’ll reach my hand optimized assembly either, imperfect as it may be.

The test code

This may look like a synthetic benchmark, in simplified C++:

complex volk_32fc_x2_dot_prod_32fc_generic(const vector<complex> &in1,
                                           const vector<complex> &in2)
{
  complex res;
  for (unsigned int i = 0; i  Continue reading

ChatGPT Explaining the Need for iSCSI CRC

People keep telling me how well large language models like ChatGPT work for them, so now and then, I give it another try, most often resulting in another disappointment¹. It might be that I suck at writing prompts², or it could be that I have a knack for looking in the wrong places³.

This time⁴ I tried to “figure out⁵” why we need iSCSI checksums if we have iSCSI running over Ethernet which already has checksums. Enjoy the (ChatGPT) circular arguments and hallucinations with plenty of platitudes and no clear answer.

ChatGPT Explaining the Need for iSCSI CRC

People keep telling me how well large language models like ChatGPT work for them, so now and then, I give it another try, most often resulting in another disappointment¹. It might be that I suck at writing prompts², or it could be that I have a knack for looking in the wrong places³.

This time⁴ I tried to “figure out⁵” why we need iSCSI checksums if we have iSCSI running over Ethernet which already has checksums. Enjoy the (ChatGPT) circular arguments and hallucinations with plenty of platitudes and no clear answer.

Worth Reading: Was MPLS TE Worth the Effort?

Bruce Davie continues documenting the tradeoffs we had to make in networking, this time with Was MPLS Traffic Engineering Worthwhile? I found this bit particularly familiar:

It wasn’t hard to make a theoretical argument that MPLS-TE could improve network performance and average link utilization, by moving traffic from congested links to uncongested ones. The hard part was proving that it would actually do a better job in practice than the more traditional methods such as using link weights and multipath routing to achieve the same ends.

Worth Reading: Was MPLS TE Worth the Effort?

Bruce Davie continues documenting the tradeoffs we had to make in networking, this time with Was MPLS Traffic Engineering Worthwhile? I found this bit particularly familiar:

It wasn’t hard to make a theoretical argument that MPLS-TE could improve network performance and average link utilization, by moving traffic from congested links to uncongested ones. The hard part was proving that it would actually do a better job in practice than the more traditional methods such as using link weights and multipath routing to achieve the same ends.

Will ChatGPT Make Tech Blogs Obsolete?

This post originally appeared in a slightly different form in the Packet Pushers’ Human Infrastructure newsletter. You can subscribe and see all back issues here. Daniel Miessler says AI-powered chatbots will be the end of tech tutorial blogs. And at first glance, his argument seems sound. Prior to the rise of chatbots and digital assistants, […]

The post Will ChatGPT Make Tech Blogs Obsolete? appeared first on Packet Pushers.

Climate Simulation Screams On The Frontier Exascale Supercomputer

We will always complain about the weather. It is part of the human condition. …

Climate Simulation Screams On The Frontier Exascale Supercomputer was written by Timothy Prickett Morgan at The Next Platform.

Heavy Networking 674: IPv6 Essentials For Network Engineers – Think Abundance, Not Scarcity

Heavy Networking 674: IPv6 Essentials For Network Engineers – Think Abundance, Not Scarcity

On today’s Heavy Networking we get into IPv6 essentials for network engineers, including how to incorporate IPv6 support in upcoming projects, how IPv6 affects NAT and subnetting, what the heck Happy Eyeballs and nibble boundaries are, and why you should approach IPv6 with a mindset of abundance not scarcity.

The post Heavy Networking 674: IPv6 Essentials For Network Engineers – Think Abundance, Not Scarcity appeared first on Packet Pushers.

Consuming Content the Way You Want

One of the true hidden gems of being a part of a big community is the ability to discuss ideas and see different perspectives. It’s one of the reasons why I enjoy working at Tech Field Day and why I’m lamenting the death spiral of Twitter. My move to Mastodon is picking up steam and I’m slowly replicating the way that I consume content and interact there but it’s very much the same way I felt about Twitter thirteen years ago. There’s promise but it needs work.

As I thought about my journey with social media and discussed it with people in the community I realized that a large part of what has me so frustrated is the way in which my experience has been co-opted into a kind of performative mess. Social media is becoming less about idea exchange and more about broadcast.

Give and Take

When I first started out on Twitter I could post things that were interesting to me. I could craft the way I posted those short updates. Did I want to be factual and dry? Or should I be more humorous and snarky? I crafted my own voice as I shared with others. My Continue reading

On the ‘net: The CCIE Shuffle Podcast

The Cisco Certified Design Expert (CCDE) exam was launched in 2007, but not many people know what the main objectives of the certification were at the time. Who better to enlighten us on some of the thought process and reasons behind the exam being created than one of the original development team? In this podcast, we are extremely humbled to be joined by networking industry legend, Russ White who spoke about his career, how he got into networking and some insights on the CCDE concept and how it came to fruition in the early 2000s!

On the ‘net: Privacy and IPv6 Renumbering

Hedge 174: Javier Antich and Cloud AI

ChatGPT has broken through the hype barrier and brought AI hype to the larger world. But what does AI mean to network engineers? We’ve talked about AI driven network management for years, and commercial products abound, but what does it really mean to move from the automation driven configuration to AI driven decision-making? Javier Antich joins Tom Ammon and Russ White for this episode of the Hedge to talk about cloud AI for network engineers.

download