> Glance over the generated assembly and make sure that it vectorized the way you wanted.
Isn't that sth you would also need to do in Fortran? IMO Julia makes this so easy with its `@code_*` macros and is one of the main reasons why I use it.
If you write Julia similar to Fortran, with explicit argument types and for loops and avoiding allocations it shouldn’t be too far off. Fortran IIRC has a few semantics which might make it more optimal in a few cases like aliasing
But indeed there are almost certainly less performance surprises in Fortran
The lawyer team of the father, "Abogados Cristianos", have a bad reputation here in Spain for suing everything they can that it's "not-conservative": from women trying to abort to comedians due to some joke. Most of their lawsuits are rejected, but when a judge accepts it they get some notority in the media.
“Abogados Cristianos” (conservative association) has the same reputation for the left-leaning than “El jueves” (far-left comedy publication) has for the right-leaning.
PhD student there. Ask me anything if you wish to know more about BSC.
Some news: the new Marenostrum 5 no longer fits in the chapel, so it has been moved to the contiguous building. But one quantum computer from the Quantum Spain project will be installed soon (in a couple of months).
Easter Egg: whenever there is an official visit, they put Gregorian choirs in the chapel.
I'm working on classical simulation of quantum computers using tensor networks. Basically trying to push the frontier of quantum advantage / supremacy from the classical side.
BSC is fun due to its interdisciplinarity. I do quantum and HPC, but I have friends working on iron deposition on the seas (i.e. climate change), nuclear fusion simulations of a tokamak, large scale scientific visualization, protein folding and synthesis, ...
The European Chips Act is also coordinated from there so there is a ton of people working on hardware design (specifically RISC-V).
This is not true. BSC had already a growing trend before the pandemy. In 2020-2021, before any EU fund, almost 200 new people were hired.
Why are you harsh? Science is a good investment and BSC has proven to be a research center where a lot of quality research is done.
Anyway, there are many plans to continue growing beyond the NextGeneration EU funds, which usually the steering committee explains on the yearly anual meeting. Keep in mind that "growing" does not always mean to grow the amount of people in a research group, but to try to create new research groups to keep it stable.
> Why are you harsh? Science is a good investment (...)
Like any high-risk investment, science is a good investment because sometimes one out of N researchers is able to have meaningful output, while the remaining N-1 barely manage to justify their existence. This might be a easy risk to take for those who take in funding, but it is also has negative tradeoffs for society in general and researchers in particular.
Think for a minute about research candidates and their careers. They are admitted as researchers fresh out of college, they spend years doing hard intellectual work on subjects with little to no relationship with industry in an environment which arguably makes you even less suited to work in industry, their livelihood in the short term becomes dependent on their institution's ability to attract funding and in the long term depends on some research institutions opening tenured positions.
Look at BSC. Do you think there will be 500 tenured positions being opened in the next decade for these researchers alone? And these are the guys already in the pipeline. 300 tenure positions per year is unthinkable. More will enter it next year, and the next year, and the next year, etc. What's your plan for these people? How do you expect them to enter the job market?
Academia is a meat grinder that's fed by graduates and spits them out, most of the time with nothing of value in terms of relevant skilsets and professional experience, and dumped onto the job market already with an age band that's incompatible with entry-level positions.
In the meantime, the money spent on these research positions could be used elsewhere. A low/mid density residential building costs around 1 million to build, and houses a dozen families or so for life. God knows Barcelona is experiencing a major housing crisis. Is this a good tradeoff?
Where is the harshness? Asking “What’s the plan when the gravy train stops?” Is a perfectly valid question.
> Science is a good investment
Some science has been a good investment. Most (especially public funded, particularly in Spain) has been a very poor investment. As an obvious example, the new accelerator CERN is pushing for will have an atrocious ROI, just like most money spent on space exploration.
We should not be taxing blue collar workers so that middle class “scientists” can call dibs on historic milestones.
Depending on the department, about half of the private investment can be up to 50%.
> Where is the harshness? Asking “What’s the plan when the gravy train stops?” Is a perfectly valid question.
I'm pointing to the other comments. I also try to answer question in the last paragraph.
> Some science has been a good investment. Most (especially public funded, particularly in Spain) has been a very poor investment.
In _general_, science is a good investment. I don't agree that there have been poor investments in Spain. Do you have numbers or reports about that?
Most cases I know are that in Spain the Big Crisis cut investment and getting stable long-term fundship is almost imposible, which makes scientists' situation very unstable. And you need long-term investments in science to make returns.
> As an obvious example, the new accelerator CERN is pushing for will have an atrocious ROI, just like most money spent on space exploration.
It's not obvious to me. The LHC and space exploration have had huge indirect returns from the technology designed around it.
An you can make up to 200% a year trading stocks! “Up to” is doing a lot of work there.
> In _general_, science is a good investment.
Citation needed. Most calculations of ROI that I’ve seen show that publicly funded science has lousy returns. They need to fuzz the numbers a lot just to make it sellable to the public.
> I don't agree that there have been poor investments in Spain. Do you have numbers or reports about that?
I’ve even worked for them, the UPC, just besides the BSC, has had historically wasted buttloads of public money.
I know of a department that blows > 1M€ of public funding every year in a project that was old news when it started, 15 years ago. It’s still ongoing. They’re not outliers by any stretch.
> Most cases I know are that in Spain the Big Crisis cut investment and getting stable long-term fundship is almost imposible, which makes scientists' situation very unstable
The number one issue is that most projects being chased are terribly obsolete. Spanish science is not underfunded, it’s just very poorly managed. I’ve personally seen managers wasting 20k€ in buying useless parts, because they could not be bothered to do the numbers first (5 minutes of back of the envelope calculations would have sufficed).
Let’s not forget that Spain has a whole ecosystem of semi-public companies (like Eurecat) whose business model is to find partners to justify getting grants. The most honest people there will tell you straight (this is an actual conversation I’ve had) : “this project isn’t going anywhere, but your company will make an easy 50k€, we’ll help you with the paperwork, and you’ll get something nice to post on Linkedin”.
> The LHC and space exploration have had huge indirect returns from the technology designed around it.
That’s the most absurd cop out people always use. “If we spend 100MM€ building this thing that barely has any uses, maybe, we might find some stuff around the way”. We are living in the era of Luck-based development, You point to a random goal and hope to make returns by stumbling into something along the way.
Even if we took all tech “developed” at CERN as a return, I’d not value it even closely to the money spent there.
If you don’t believe me, you can ask other people who have worked there, the amount of wasted money is ludicrous. Just as an example, my team designed a custom ASIC for the LHCb. There were literally 4 identical ASICS being designed in different teams.
And let’s not get on the promises made to sell these projects. Just check the writings and videos of Sabine Hossenfelder, and you’ll start to get a grasp of the grift, and IMO, she is too diplomatic about it.
I'm on that project myself, we are also dealing with simulating circuits, buuut I feel like we could improve our approach somewhat lol. For example, we had a bunch of trouble trying to speed things up, I immediately thought of trying to make most things run on a GPU but we quickly found out that our circuits just have too few qubits to be parallelized decently.
More importantly, I am not very optimistic (to say the least) about the short- to medium-term real-world applications of quantum circuits we are looking into (we do time series classification, which is quite removed from other domains which work better on these computers), and I got the same feeling from the literature. Should I be feeling differently?
Yes, but you must contact through a form or get someone from inside to show it to you. Unfortunately, I'm out of town til next week so I cannot offer it myself. Check out https://www.bsc.es/discover-bsc/visit-our-supercomputer
If you want to visit some other nerdy places, I suggest you to visit:
- CaixaForum: cool science museum
- Observatori Fabra: a very old telescope (no longer used for science) but has cool views of the city
These days there is a very cool interactive AI exposition at CCCB. I really recommend it.
If the guide is not available just DM me one or two days before you arrive. Any BSC staff can give a tour to a small group. But it is necessary to schedule ahead and avoid events/other tours.
You can also visit the luminous click on Via Laietana. It is old, free (on the sidewalk), but interesting.
You can enter any library to browse. Abello library in Les Corts hosts the Asimov Foundation collection. I think the Sagrada Familia library has the scientific illustration collection, etc.
I remember seeing a really old elevator in the Gòtic neighborhood... I can't remember the name of the museum where it was but it was at the museum entrance (you can just walk in and have a look).
You can also walk in most campi like UPC, UB, etc. They have some exhibitions and also panels with history of the uni and faculties. As well as libraries too.
The quantum computer is corruption. They created an spin-off where some management guys have shares on the company. Then launched a procurement. Only the spin-off company went for the procurement and won.
Sounds about right here in Europe. At least they have something to show for it this time, I guess? I mean, assuming they created the quantum computer with the EU funds.
wtf are you talking about? I know people from both sides and can assure you that the whole process was done transparently and legally. The external board which selected the winner proposal did a blind review (they only knew the proposals, not who made those proposals).
It's true that Qilimanjaro is a spin-off of BSC, but they are the only company in Spain that can fabricate quantum chips. Furthermore, the chips are fabricated by Quantware and IQM, while Qilimanjaro provides the testing, calibration and service. One of the requirements from the European Funds is that the providers must be European companies. There are not many fully european companies that fabricate digital quantum chips, I'm afraid.
If you have doubts about the legality of the process, I welcome you to contact the coordinator of the project.
So they knew each other. And instead of doing this from a BSC salary they launched an spin-off and then sold their work to the BSC being part of BSC while he was working 50% of his time for BSC and therefore being a public employee and having an incompatibility.
You obviously do not understand tenders and also have a bad faith in addressing this message towards me (García-Ripoll) and Artur.
First of all, this tender only received one application. There could have been other applicants as it happened in Galicia's quantum computer tender, but there were none.
Second, this application was done by a collective of companies of which Qilimanjaro is only one more participant, responsible for calibration and integration.
Third, my role as scientist in this evaluation was only to certify that the specifications of the tender proposal follow the scientific requirements of the call, and whether any parameter is below or above the requested standards. That is tickboxing essentially and with only one applicant this does not influence the outcome.
So I am asking you to retract these accusations as I am concerned, or follow suit if you have so much evidence of wrongdoing.
YOU obviously don't understand tenders. You can not create your own company inside a public company and then give then money by running a procurement.
1. They knew about the tender before it was public (illegal)
2. Conditions for procurement were adjusted for this company (illegal)
3. Qilimanjaro needs experience and at least 3 contracts of the same nature fullfilled correctly. (Irregular)
4. They did NOT go in UTE (temporary union of companies) and the subcontracting was NOT declared (illegal)
I have nothing to do with this fight, other than being Spanish myself. But it deeply bores me that it has occurred.
I don't care who is right, truly. What bothers me is that this isn't the appropriate forum, the unbearable idealism of the accusation, the unnecessary defense, the veiled threats, etc.
What really burns me is that we are always like this… Me tuvo que tocar un país de patio de colegio.
Before answering, I want to be honest and state that Artur is my supervisor, so I might be biased. Now that I've said so...
1. Artur doesn't have an incompatibility, he asked for permission to the ministry and he works his daily 8h (sometimes more) at BSC. After that he works at Qilimanjaro.
2. Juan José García Ripoll has nothing to do with Qilimanjaro. He has his own company (Inspiration-Q).
3. Juan José was a collaborator many years ago.
4. You would be amazed to see how small is the quantum world. I've already met many of the top researchers just by moving around, because everyone knows each other. So by that logic, no one would be free of conflict. Of course they knew each other. Every one knows Juan José. Many measurements are taken to prevent favor treatment in these situations.
6. We removed the website because it was no longer being updated.
I've seen from your comments that you have been falsely accusing without any info from the inside and that all you want is to harm all the effort done by these amazing people to make progress in science in Spain. So I will stop answering your comments. I don't know where this hate comes from, but I invite you to introspect.
If anyone reading the comments gets dubious about the process, I invite them to visit the link provided above to check all the documents of the procurement and to get in contact with the project coordinator. The process was done cleanly and transparently, and as any legal public project, it can be checked and verified.
Ps.: It made me laugh to check all you stalked about Artur. I think nobody has done it before. Will show it tomorrow to Artur jajaja.
When you work for BSC you sign a non-compete agreement saying that you won't be selling all you learn from BSC to BSC:
Point 5.2: prohibición de aprovechamiento en beneficio propio o de terceros. Queda prohibido que el trabajador, sin consentimiento del responsable de
BSC, transmita, reproduzca, destruya o utilice información referenciada en el apartado anterior para uso distinto al normal de trabajo.
Del mismo modo se prohíbe estrictamente atodos los empleados y colaboradores de BSC el uso de dicha información en beneficio propio o de terceros.
Point 5.3: Cláusula de exclusividad. Queda prohibido que el trabajador realice proyectos basados en conocimientos del BSC para terceros fuera del ámbito
laboral, o por encargo directo de terceras personas. Una vez desvinculado del BSC, queda prohibido que el trabajador aplique conocimientos propiedad
del BSC o de colaboradores del BSC, con fines lucrativos, o al servicio de terceras empresas que los exploten económicamente.
We were running with the same problem (supercomputer with clusters of different architecture and no outgoing connections permitted) and so we created "pypickup" [1,2]. nice to see that we came with similar solutions! I have some questions:
1. is the directory of packages you create compatible with the PEP 503? (so I can use `--index-url file://PATH_TO_LOCAL_CACHE` flat with pip and it should work)
2. is there some filtering mechanism? e.g. we are not interested in non-release versions ("dev" versions, "rc" versions, "post" versions, ...)
3. I guess that the way morgan resolves dependencies is by manually parsing files like "pyproject.toml" or "requirements.txt" and it does not ask the build-system for the dependencies. if so...
- does "morgan" detect build-dependencies?
- which build-systems are compatible?
- is "morgan" capable of detecting more complex dependency specifications? e.g. "oldest-supported-numpy" which is used by "spicy" has dependency strings like the following: numpy==1.19.2; python_version=='3.8' and platform_machine=='aarch64' and platform_python_implementation != 'PyPy'
Too bad your project didn't come up in any of my searches while researching this problem. Probably because it doesn't use the word "mirror" at all :)
As for your questions:
1. I don't see any mention of directory structures in PEP 503. The Morgan server does implement PEP 503 though. In any case, I tried installing now straight from the directory and it didn't work. Are you sure you meant PEP 503?
2. Where Morgan differs from pypickup, as I can see, is that it interprets requirement strings as per PEP 508 (e.g. "requests>=2.40.0; python_version < '3.8'") instead of providing a command such as `pypickup add requests`. For every requirement string, it looks for the latest version in PyPI that satisfies it, and downloads that version. You can filter _in_ the requirement strings, other than that Morgan doesn't have any specific handling of dev/rc/etc.
3. Morgan detects and downloads the build system based either on the [build-system] section of pyproject.toml, or the setup_requires.txt file (from setuptools). These are the sources currently supported. It doesn't actually care what the build system is, it simply attempts to find where it is defined and download it as well.
As for complex dependency specifications, yes, they are supported and honored (Morgan relies on the "packaging" library to properly evaluate those). By the way, I recently moved from Poetry to Hatch for managing the Morgan project itself specifically because I got fed up with Poetry not honoring those specifications, and trying to download completely irrelevant packages.
Well, we first named it "pypi-cache" but there is a package named "pypicache" from the year 2007 and we had to rename it. We always thought of it as a "cache" rather than a "mirror"... but yes, "mirror" is more appropriate. Btw we released it just 1 week ago which is also maybe why you did not find it.
1. Well, the flag "--index-url" explicitly says that "... should point to a repository compliant with PEP 503 (the simple repository API) or a local directory laid out in the same format". PEP 503 defines the directory structure where there is a folder per package, an "index.html" on the root with a link to each package and *an "index.html" in each package folder that has a link per available file*.
URLs are not limited to "https", they can also be relative paths. So the trick we do is to download the file to the folder of the package and add an anchor to that file in the "index.html" of the package. For example,
But we download it and write,
<a href="./numpy-1.22.4.zip" data-requires-python=">=3.8">numpy-1.22.4.zip</a>
This is specially important for us because we cannot setup any kind of server.
2. Okay nice. Yep, we thought that parsing would be more difficult and that relying on parsing would be problematic due to the different build-systems and that many packages still do not have the "pyproject.toml" file. We opted for a manual approach in which you do "pypickup add" until you have no more "dependency missing" errors. Your approach looks much better to me, but like you said is limited to "pyproject.toml" and "setuptools" right now.
Btw, does it also downloads extra dependencies?
3. Nice. I also stopped using Poetry for things like that, but now I manually write my "pyproject.toml" with "setuptools".
I like the idea on trying to parse the dependencies. I will probably try something but since we download all files (filtering some of them), it would be more costly. Maybe in some weeks when I'm more free.
Ahh, I get it, it needs index.html files. I can easily implement this, but I actually did want the server because I wanted it to be easily accessible from multiple machines, I also wanted to implement the JSON API, and also want (in an upcoming version) to allow uploading private packages to the mirror.
As for extra dependencies, yes, they will be mirrored, but only if relevant, i.e. if they are included in a requirement string (be it a direct requirement or a dependency of a dependency).
Isn't that sth you would also need to do in Fortran? IMO Julia makes this so easy with its `@code_*` macros and is one of the main reasons why I use it.