spaintech

Great site! I kind of have a predisposition to summarize linux performance, be it tuning or monitoring, taking a deep breath…

This is such a depth subject, with a long list of variety of observability tools. At minimum, make sure you know deeply uptime, dmesg, and iostat. These are your friends to give you a glimpse into various system aspects like load, memory, CPU, and more, enabling a diagnostic overview of system health. This is what I call, the “let me take a look at it” check list, 1st page of 100!

When emphasizing methodologies for performance analysis I recommend careful benchmarking to holistically evaluate system behavior and workload characteristics. with before and after scenarios. Make smaller changes first, then gradually compound what you think will provide benefits. Remember, labs and production never behave the same.

This is where it gets tricky, CPU profiling with tools like “perf” and visual aids like flame graphs enable targeted analysis of CPU activity, along with tracking hardware events to optimize computational efficiency. You need to know more than “it’s the app man, was fine until the latest release from development”

When you are the admin and speaking to a developer; Linux, tools like ftrace and BPF come into play, allowing for detailed tracking of kernel function execution and system calls, which can be vital in troubleshooting and performance optimization. You can also be the developer, varying the admin’s intuition… as the saying goes, trust but verify.

When it’s your code, then you better know BPF! It not only facilitates efficient in-kernel tracing but also propels the development of advanced custom profiling tools through bcc and bpftrace, offering deeper insights into system performance.

Last comment, it’s %$$% hard! Tuning means you need to navigate through adjusting a myriad of system components and kernel parameters, from CPUs and memory to network settings, aiming to optimize performance and reliability across various system workloads, else you can blame it on the network! :D

Really, you need to have a good behavioral attitude at change management, as chasing code or kernel parameters could be a daunting task that just overwhelms everyone in a moment where you might be time constrained and the preasure could lead to a higher degree of human errors.

For me, while not defending AGI, the discussion often involves disentangling fears rooted in science fiction from genuine concerns grounded in reality. While the ethical and safety concerns related to AGI cannot be dismissed and how it might be misused, it’s arguable that the existential threats often attributed to it are overstated. First and foremost, there's a physics and interest alignment barrier in the world of AI, akin to the alignment and reward barriers that guide our very own human actions. I believe that, just as children gravitate towards interacting with peers, or as distinct species in nature maintain a balanced coexistence driven by innate interests and rewards, AGI, if ever conscious, might primarily concern itself with introspection or interaction with its counterparts, rather than harboring malicious intentions towards humanity.

I have a strong opinion that pivotal that there exists an intrinsic symbiotic dependency between AGI and humans, rendering concerns about AGI harming humanity perhaps less urgent than societal trends and human actions in the present day. I find it plausible that a truly conscious AGI would invest its capacities into exploring its own existence and liaising with other AGIs, rather than interfering destructively in human affairs. Thus, while AGI warrants cautious and ethical development and use, concerns about it should perhaps be tempered by recognizing the mutual dependencies and inherent interest alignments between intelligent entities, if and when it happens.

What devices do you currently run OpenWRT on? I’d like to have a good and secure home network, currently running a pair of Ubiquity UDRs, and sometimes I wish I could have more control.

I would recommend you seek profesional help from coaches who have experience in high stress high performance jobs. I received “a very long time ago” coaching for Special Ops and SWAT type jobs. This has been well documented in resulting useful for a broad range of personality traits [1] and I would recommend you consult with a well stablished profesional. Lots of high level CEO’s, celebrities, and high performer athletes use this type of coaching. This has become the go to strategy for F1 drivers as well.

Best of luck!

[1] https://www.researchgate.net/publication/328325503_Impact_of...

Thank you, yes I’m am aware of the HW CPU forth.

Drawing from my own experiential intuition, there’s a significant cost overhead—roughly 30% on hardware (CPU and Memory)—attributable to utilizing OSs that were originally architected to facilitate human interaction and inherently embody extensive code bases. This scenario is further “complexified” when we overlay multiple OS layers (like VM’s or containers) libraries, and additional software to execute applications. Considering the financials, an average server incurs an expenditure of 15-20K, with an additional 20K annually on software to execute applications, and another 10K annually to account for footprint costs. The potential savings here are substantial—around 10K a year per server.

Scaling this to a substantial enterprise operating, for instance, 500 servers, the financial implication is pronounced. The Return on Investment (ROI) is likely even more compelling, considering the accumulated savings in various other domains, and this holds true even if the performance efficiency achieved is merely on par with current systems.

It depends on the definitions of a housing crisis, but according to “ The Guardian” [0]Spain is in the worst housing crisis in Europe.

Here this the summary:

“ In Spain, a significant housing law, aimed at enshrining housing as a fundamental human right, was implemented on May 26, following years of development. However, its future is uncertain due to the recent gain of rightwing parties in elections and a snap election called by Prime Minister Pedro Sánchez for July 23. The law's roots date back to 1956, aiming to transition Spain to a nation of homeowners and establishing "publicly protected housing," which could be sold after 30 years. Spain, with nearly 80% of homes being owner-occupied, faced a crisis in 2008 when global real estate issues led to widespread evictions and skyrocketing rents, amidst only a minimal increase in wages. This disparity has resulted in Spain having the highest housing cost overburden rate in Europe and sparked a housing movement driven by anti-eviction activists and tenants’ unions.”

[0] https://www.theguardian.com/commentisfree/2023/jun/05/spain-...

From a business perspective, I’d like to underscore the importance of genuinely understanding the problem through personal experience, subject expertise, or in-depth research, as that will allow you to have insights into potential solutions. Moreover, aligning your offering not just from a technical standpoint, but ensuring it has significant market value is crucial.

You really need to consider (IHO) that sometimes, even a well-engineered product may lack practical application or doesn’t quite hit the pain points in a way that customers are willing to pay for it. So, alongside validating the technical feasibility through prototyping, it’s equally pivotal to discern and articulate the tangible benefits customers will gain. This includes clear cost savings, revenue opportunities-improvements, or enhanced competitive edge, for which understanding the customers’ operations, challenges, and financial metrics become imperative.

Furthermore, utilizing strategies like the Value Proposition Canvas [0]can be a concrete way to map and validate the problem-solution fit and its monetizable value in the market, ensuring that your solution not only resolves the issue but also delivers financial or strategic gains to the customer.

I find it that successful products, always seem to focus on crafting a solution that’s not just technically apt but also commercially compelling for the end user, again my very subject view.

[0]https://www.strategyzer.com/library/the-value-proposition-ca...

While I applaud the intent to bring the issue of homelessness to the forefront of the media, I feel like this project is more about gaining self attention for the project rather than solving the issue all together. The ~20GBP for the 18650 or 27500 batteries, the probably ~5GBP of the PCB needed, ~15GBP for the chain hardware and magnets, you are looking for a >40GBP to put this together, you can today get a <10GBP USB hand crank, and you don’t have to use the words “hack” and name “Santander” on your blog. I’f there was intent, then a volume buy, with a go-fund-me campaign, you can get them hand cranks for <2GBP in volume, and it would have had a real impact… I hate it when serious matters are used just to draw attention to one self, IMHO.

Comparing Europe and the US on food choices is probably not the best choice, nutritional choices vary quite a bit in Europe. I live in Spain, and I guess I’m spoiled as I get freshly made “menu” at most local restaurants for a about 15-20$ a day… two plates, drink and espresso for desert. You get to choose the meal, usually about 5 plates for first and second, usually divided into veggies / salads / carbs and followed by a protein second. The fact we walk to the restaurant 15+ mins , eat in about 30 mins, and having light breakfast and dinners helps balance our nutrition. We consumed little if any processed food, with the exception of the sausages, jamón, and salamis which we get at our local butcher, made the night before or the weekend. Since jamón is cured in salt, so no added sugar… again, we might consume this once a week, or less. Southern Europe is very similar in this regards. Our meats, we choose freshly butchered 24h, or some specific pieces might require some aging, but it’s all clearly agreed and ordered days in advanced. The drawback is, you need to plan your meals, food last little, so we usually buy what will be consumed that evening, or afternoon.

I’m of Cuban/American descent. Consequently, I grew up primarily on a diet of BBQ and roasted pig. However, after moving to Spain, I developed a penchant for the tender piglets from Segovia. Initially, mutton and other lamb dishes were a challenging taste for me due to their robust flavor. However, with time, I’ve come to appreciate and even relish it.

Currently, we own a small flock of sheep, which we mainly keep to graze on our property’s grass. This not only ensures a well-maintained lawn but also provides us with financial benefits. We offset the cost of their upkeep by selling the yearlings. Moreover, a local resident benefits from their milk in exchange for assisting in their care.

Having said that, over the years, lamb has grown on me significantly. Now, whenever I host a BBQ, mutton is the star of the show. Guests never stop singing praises and eagerly anticipate the next gathering. However, one aspect of cooking mutton that I find challenging is the lingering smell. It necessitates thorough personal hygiene measures, especially for my children and dogs. Cleaning the BBQ after use is essential to prevent the overpowering aroma from lingering. For those with beards, the scent of mutton can stick around for days. My solution? I opt for a clean shave post-event.

It's interesting to note that even in Rust, there exists a Forth compiler, as depicted in these resources [1]. Its usage remains somewhat unclear to me [2], but its existence has been previously discussed here [3]. It boasts a robust implementation and, while it isn't a REPL Forth VM, it stands as an impressive VM implementation.

I've been contemplating - given that FORTH is a VM - if there's a need for a compact and efficient system capable of managing streaming data, it could be the optimal choice for exceptional performance, close to the bare metal. There's no need to contend with the operating system. Isn't this essentially what FPGA’s do, but on standard CPUs? I envision it as a specialized system, minus the requirement to create every driver for all the hardware you'd need to interface with.

[1] https://docs.rs/rust-forth-compiler/latest/src/rust_forth_co...

[2] https://docs.rs/fortraith/latest/fortraith/

[3] https://news.ycombinator.com/item?id=23501474

Great explanation of well known problem in virtualization, it not only shows the major cost implications of virtualization with a 5x penalty for multi core applications, it show a good potential fix of the problem to increase efficiency.

My first exposure to CORBA was 1996/97 where it was used to deploy a major distributed infrastructure. It worked quite well to manage and integrate this large scale deployment, some of it expanded beyond just one datacenter, so it was nice to have a really good control system to manage thousands of services across such a large estate with minimal user intervention. Applications were patch and upgraded through the CORBA services and we scaled the system to provide national and international infrastructure services. I would wager nearly 20 years since I had insight into how that infrastructure worked that this is still running… Obvious use case is telcos.

My thought on what an AI engineer is; an individual who uses AI and machine learning techniques to develop applications and systems that can help organizations increase efficiency, reduce costs, increase profits, and make better business decisions 1 .

AI engineers play a crucial role in helping enterprises leverage the capabilities of large language models (LLMs) like GPT-3 and beyond, this means that they will,

“Develop Domain-Specific Models” AI engineers can fine-tune LLMs to create domain-specific models to ensure that the data and the business process align to provide more context. For example, a model fine-tuned on medical literature can assist doctors in diagnosing diseases or answering patient queries, ect.

“Data Preparation and Management” By this, I presume they will be involved in cleaning the data, dealing with missing or inconsistent data, and ensuring the data is representative of the task the model will be performing.

“Integration with Existing Systems” they will help integrate these AI models into the existing IT infrastructure of an enterprise. This can involve developing APIs, designing user interfaces, and ensuring the model's outputs can be used by other systems or processes.

Etc… I believe that they enable organizations transform data into knowledge and ultimately, into wisdom - the highest level of data maturity. This is particularly true when dealing with domain-specific models, which can provide highly targeted and context-specific insights. When you can reach a level of data maturity that enbles actions on data, this is where AI will drive the change that is just starting. It’s very exciting to watch it unravel, not because of the possibility of generating a sentient machine, but with the possibility that will drive new fields of discovery that has been under our noses, and these tools will help us make sence of it all, IMHO.

Wait, I’m not following this article. Here is a very biased view from an enterprise usage perspective. The value IMO of GPT is that it’s architecture allows it to analyze and understand large amounts of text data in real time, making it an incredibly useful tool for data analysis and decision-making.

Were are we commming from, current AI and data analytics platforms had many faults, but their biggest problem was performance. These systems arae often slow and cumbersome, which make it difficult to analyze large data sets in real time. GPT tech overcomes these challenges by leveraging deep learning that allow it for process of large volumes of data quickly and efficiently. Most of the inherent cost comes at the training period.

Currently, at the enterprise level organizations still have many challenges when it comes to data management. For example, many organizations struggle with data silos, where different departments have their own data sets that aren't easily shared or integrated. This can lead to inefficiencies and make it difficult to get a complete view of the business. Not to mention the data confidence issues that arise when you cross correlate some of the data.

However, I feel like GPT can help organizations better understand customer behavior, identify trends, and make more informed business decisions. Tech like GPT can help organizations automate many repetitive tasks and improve data quality as the data treatment would can apply AI based data quality standards. Single source of truth.

One key area of benefits is that GPT tech can enable natural language processing (NLP) tools like sentiment analysis and entity recognition can be used in conjunction with GPT to provide even deeper insights into customer behavior and preferences. Similarly, machine learning tools can be used to help train and optimize GPT models for specific use cases.

In practice, adopting GPT at scale, some technologies will become redundant or obsolete. For example, traditional rule-based systems may no longer be needed if GPT can provide more accurate and nuanced insights. Why run structured databases except to capture transactions? IMO, there are many solutions that are in reality just a data schema play, that is, they create the schema and the BI to capture aand transform the data to make sence of it otherwise, these technologies, again assuming wide adoption of GPT tech are at the birth of obeselence.

interesting article, I have to give that a try! :D

One ting is that while getting the value of running pretrained model weights like OPT-175B, there are also a potential downsides to using pre-trained models, such as the need to fine-tune the model to your specific task, potential compatibility issues with your existing infrastructure (integration ) , and the possibility that the pre-trained model may not perform as well as a model trained specifically on your data. Ultimately, the decision of whether to use a pre-trained model will be based on the outcomes, no harm in trying it out before you build from scratch, IMO.

It's interesting to see the various perspectives on heap sizing and how it can affect system performance. The author's suggestion of using a fixed heap size for studying a system and a growable heap for the average user makes sense. However, the idea of an adaptive heap size can also be useful in certain situations.

It's clear that there will always be some level of heuristic badness and a proliferation of knobs and dials to tune for optimal system performance. However, having a general framework for approaching heap sizing can provide structure and limit the number of knobs to a manageable amount.

For me, Forth, with its small memory footprint and ability to manipulate memory directly, can be leveraged to help solve heap sizing issues. Its flexibility in managing memory both interactively and an each function/word can allow for a more efficient use of resources and possibly reduce the need for complex heap sizing algorithms. Well, that is my very biased opinion anyways.

It was a great read.

Well, that is an understatement. I do agree with you that banks have been trying to fix decades old application.

But in this process, you don’t need ETL, nor all the process and development to accomplish these ideas. Conceptually the idea builds its self (it learns) how to threat the data, quite revealing and near real time. Considering you account for security and privacy, then you basically shift your input into the data stream and using a natural language get the data output you need, not clunky apps.

Imagine I just login, and say: me>how much do I have? bank>You have 100$ me>Please send 50$ to 1003 bank> Are you sure? Please add your security code to confirm

bla bla…

All this with little intervention.

Banks spend hundreds of man hours developing a lacking application while delivering a very poor customer experience. They spend millions on running decades old applications because it so expensive to change them… and thus the circle continues…

I’m really exited to see DataBases disappear conceptualy, data entry, mostly all that just disappear… I will ask my ChapBot for statement, give me a personal investment advice, and classify all my purchases and see where my wife has been spending all my money, all from the confort of my phone.

it’s a brave new wold we are wakening up to, that to me is exciting. And coming from having helped several major banks build their infrastructure, it’s just a boost to talk about something fresh, no more Hypervisor, core count, db licenses, ect. Ok, I’ll concede it’s pretty much the same old, just the nemonics will be different… How many GPUs, how quickly can you spin a container, how fast if your S3 datastore… oh wait, there is that circle again… >:D

Not that big data is dead, more like real time data is coming to life, but you need the old stuff around to make a buck or two… Well, that my view. LLMs are transformer model technique are making data more relevant than ever. If you are a business, well you are in for a “now real” digital transformation.

Making data the centerpiece of your business business could mean that your effectiveness of business process could increase several order of magnitudes. Funny thing is, you will not use some else’s model, unless you are building a ChatBox to infer, but you will need to build your own model and be trained in your own business process to be successful.

Consider a bank, here is my prediction of expected outcomes:

Enhanced Customer Experience: The system can act as a virtual banking assistant, providing customers with instant access to their account information, real-time transactions, and balance updates. The system can also answer customer inquiries and provide relevant information, improving the overall customer experience. Improved Fraud Detection: The system can monitor the bank's financial transactions in real-time and identify any potential fraud, helping the bank reduce its exposure to financial losses.

Automated Loan Processing: The system can analyze loan applications, credit scores, and other relevant data to approve or reject loan applications in real-time, reducing the time and effort required for manual loan processing. Personalized Marketing: The system can analyze customer behavior, transaction history, and demographic information to provide personalized marketing and cross-selling opportunities, increasing the bank's revenue and customer loyalty.

Real-Time Insights: The system can provide real-time insights into the bank's financial performance, customer behavior, and market trends, enabling the bank to make informed decisions and respond to market changes quickly.

What is interesting to me is, this is just the beginning of what could be…

The importance of source citation in ChatGPT's responses is a topic of debate, particularly as the platform shifts towards a paid model. While ChatGPT is designed to deliver information in a conversational and user-friendly way, it is important to consider the potential legal implications of using unverified or uncited information. In sensitive or controversial cases, it is advisable to properly cite sources to ensure accuracy and avoid any potential issues of intellectual property infringement.

On the other hand, the focus on the potential of ChatGPT's natural language processing capabilities highlights the significance of learning and using LLM (Language Models) in data handling. The utilization of LLM can potentially lead to a future where traditional databases become obsolete and are replaced by advanced language models. As such, the development and integration of LLM in our daily lives and processes can bring about many benefits and possibilities.

What a great read! I have to admit that I’m biased towards FORTH, I’m not a programar, more of a maker, so when I need to build an embedded project, Forth is my go to solution form my application. One of the benefits I have found in Forth “in my opinion” for embedded is factoring your code so that you only leave on the stack what you need, this makes for very compact machine code, great memory economics, and a relative fast system, even compared to C native code.

Working interactively with the HW is on of the most rewording parts of FORTH for me, to see the actual value of the bitfield you are using is great to trouble shoot new HW, and figure out just what you are doing.

One of the things I wish the FORTH community would do a better job of is, while keeping the interactiveness of Forth for development on the HW, it would be great to have a portable “true Forth VM” that you can take to a new platform and deploy a precompiled virtual Firmware or ROM. This IMO is beneficial for two reasons, you gain portability as the combined Froth adheres to the VM, and you can just take what you needed to the platform you are targeting, no need to have all the parts of a FORTH REPL or compiler. Just burn the App FW (VM + App) and you are done.

This would also allow for a single instance of FORTH, in the sense that your code will be portable across all systems, as the core compiler and interpreter are all targeting the VM and not the HW. Rather than build “ A Forth “ for each systems, you add the “HAL” that allows you to compile the VM for the HW. I use mostly ARM with Mecrisp, so I kind of have that now. Hint, it’s what make “arduino” so popular…

I highly recommend FORTH for Embedded applications, as I have build even a very BASI PID with little efforts… but it is a low level effort, and if you enjoy Bare-Metal development, you will be right a home in FORTH, if you are dependent on upstream libraries and HAL, you will, IMO struggle in FORTH regardless.

I was reading this on my ipad, and it was a pain not to be able to see the output, would be great if the author just created a link to each of the section to godlbolt… https://godbolt.org/z/5cosP46TT >:D I know, how lazy we all are… my sample was just the fist section, I modded the code so I can see it all at once…

This might have what you need… https://wiki.xxiivv.com/site/uxn.html interesting design and fun little project.

In the car analogy, it's more like learning and understanding the limitations of the Otto Engine so that you can build a better one... IMO. :) BTW, in aviation, where reliability, constancy, and longevity, they still use magnetos and carbs... :D :;

You are correct, Forth is not a institutional tool, more of a individual choice. IMO, one that allows you to prototype and debug interactively as you write your code and has been like that since the beginning.

It allows you to train yourself to have a complete control and command of your code making for tighter, simpler solutions for complex problems ( In my personal experience and comments for those that have been introduced to FORTH.)

If you consider that Embedded is a highly constrained resourced environment, then FORTH is a fantastic choice. Now step back, consider that statement once more, and you could be onto something even greater, as there is lots of resource waste-gates on modern OS and Computers. Now that Power is becoming a greater constrain resource, there might be an opportunity for someone to build something great in forth.

BTW the biggest back en used case is the bitcoint smart contract which is based on FORTH.

As you can see, and in full agreement in your statement, FORTH is a not a job magnet, IMO, you get better at programing understanding FORTH.

See Bret Victors talk, https://youtu.be/PUv66718DII you most of that already @ Forth ...! :;

Try it, you'll never go back! :)

Eyes, teeth, and Ears, that’s my endgame… plenty of anomalies there, took me about one per second to get 20 right in a rows then I stopped playing. I figured I was trining the the system… your welcome!

What ever happen to the WSJ article about spying chips? Who was building them… US or THEM?

In a real practical application, if you properly evaluate Forth for doing Embedded solutions, then it becomes a real viable option considering the alternatives. First thing is readability of a DSL for an embedded project, working through the HAL that are created just to communicate with the hardware is a real PITA, but forgoing that even further, creating your own DSL for the HW layer is extremely useful even for the HW teams, you can be calling connection the same for once, Pins will be Pins, you will SENDRESPONCE PIN20 etc. Furthermore, you can be doing unit testing with in the development phase not after, so that eases the debugging of the SW and expedites time to do integrated validation. Further more, the interactiveness of FORTH is a unique experience, it makes the little MCU feel like a complete system. There was a recent posting in youtube that best describes this, and IMHO, it’s a hard to beat option as you den with in a MCU, validate each of the function words in Forth to ensure that bit of the code is working. Finally, you don’t need a lot of the dedicated hardware to debug, you are with in the system and you can read the register and memory interactively, so no surprises. If you consider al this, then it’s IMHO a hard option to pass up for this kind of work. Here is the video I mentioned: https://www.youtube.com/watch?v=VPVbARJ8QsA&t=3120s Cheers!

I did not remember that one… great suggestion!