ned Productions Consulting


Technology musings by Niall Douglas
ned Productions Consulting
(an expert advice and services company based in Ireland)


Saturday 11th November 2017 10.44am

#meetingcpp All packed up and ready to check out, though I'll probably take lunch with the conference before heading to the airport. It's been a good conference, a little too short, but so slickly run you wouldn't think its youth. I'd even rank it second best after the ACCU conference, and ahead of CppCon despite the latter's prestige. The availability of 90 minute slots is a huge winner, plus all meals provided and a free bar each night is unique.

Tuesday 7th November 2017 9.18pm

#meetingcpp This picture below doesn't do it at all justice, but the bar for this C++ conference is by far the best of the global C++ conferences. We are about as high as any building in Berlin, and it's a great view to have drinks to. Bar also has by far the best selection of booze of any of the conferences too, they have a many page menu and though the most expensive apart from Aspen, I can see some good discussions happening here next few nights.

Tuesday 7th November 2017 12.25pm

#meetingcpp Just waiting for the flight to Berlin for the Meeting C++ conference. Due to the rail strikes today, traveled up to Dublin last night and stayed in a hotel. I forget how badly I sleep first night in a hotel! Thus feeling pretty ropey, and looking forward to napping on the plane. And I'll surely sleep well tonight at least.

Thursday 26th October 2017 5.59pm

Looks like I'll be having to do a 12 month onsite stint in Dublin as there is zero remote work going in C++ all this year, and with Brexit and Trump the lack of remote work in C++ may persist for a long time. I had to take a 20% discount in day rate as well, there are almost zero contracts in C++ anywhere in Ireland this year, so rates are well depressed.

Tuesday 24th October 2017 6.00am

Link shared: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2017/p0779r0.pdf And my second sort-of-academic paper, "P0779R0: Proposing operator try() (with added native C++ macro functions!)" sent to WG21, the international standards body for C++.

Tuesday 24th October 2017 5.59am

Link shared: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2017/p0762r0.pdf My first new sort-of-academic paper written in some years: "P0762R0: Concerns about expected<T, E> from the Boost.Outcome peer review" which is a paper submitted to ISO WG21, the international standards body for C++.

Tuesday 19th September 2017 11.58am

Link shared: http://indeed.ie The contract market in C++ is so bad in Ireland right now that not one single contract in C++ is on indeed.ie (a jobs site aggregator) in the "new roles" section! I haven't seen it this bad since the depths of the financial crash :( Where are all the finance roles coming out of London through Brexit induced relocation? Or is it that Brexit uncertainty is sinking all boats?

Tuesday 12th September 2017 1.16am

Been making great use of my time without employment working on AFIO, my proposed low latency file i/o library for standardisation. I wrote a toy transactional key-value store with AFIO permitting up to 48 concurrent writers, and here is how it performs for 128 bit keys and 1Kb values with 1 thread: 1Kb values Windows with NTFS, no integrity, no durability, mmaps: Inserting 1M key-value pairs … Inserted at 518403 items per sec Retrieving 1M key-value pairs … Fetched at 2192982 items per sec 1Kb values Linux with ext4, no integrity, no durability, mmaps: Inserting 1M key-value pairs … Inserted at 695894 items per sec Retrieving 1M key-value pairs … Fetched at 5988023 items per sec To put those figures in context, LMDB which is considered one of the fastest key-value stores around achieves about 1.4 million fetches/sec with four threads.

Friday 11th August 2017 4.18am

You might remember that last year I spent lots of money on a MacBook Pro for one very specific reason: its NVMe SSD flash hard drive. Despite its supposed state-of-the-art performance, the various Windows benchmarking tools never really showed it to be particularly all that fast. Sure, with ridiculous queue depths and threads thrown at it then you could hit 3Gb/sec, but random 4Kb @ QD1 benchmarking always showed pretty poor results, maybe 2,500 IOPS (13Mb/sec) which is about a 400us average response time.

Friday 7th July 2017 3.30pm

My ZFS-on-Linux public facing server managed 289 days of trouble free uptime before today exhibiting massive i/o delay (see below) making it unusable without rebooting. That's the best it's ever achieved to date. Really shows how ZFS-on-Linux is becoming actually viable for reliable service provision, and as bug fixes land it's getting increasingly ever better. <br />top - 16:26:58 up 289 days, 23:27, 1 user, load average: 3.50, 2.19, 1.60<br />Tasks: 729 total, 1 running, 728 sleeping, 0 stopped, 0 zombie<br />%Cpu(s): 0.4 us, 0.8 sy, 0.0 ni, 86.5 id, 12.3 wa, 0.0 hi, 0.0 si, 0.0 st<br />KiB Mem: 16420640 total, 9778848 used, 6641792 free, 464 buffers<br />KiB Swap: 4194300 total, 0 used, 4194300 free.

Archive