342 karma · joined April 13, 2011
With GNS3, it was possible to set up complex network topologies of dynamips routers that would have cost thousands of dollars in real hardware without losing any of the realism as these were essentially virtualised routers.
I haven't kept up with the progress here but Cisco certification may not be the best example.
While anti-retroviral therapy for HIV/AID does not cure the disease, it has transformed a diagnosis that would otherwise have been a death sentence into a manageable illness.
If a "first order" treatment for cancer only extends lives by a few years or even decades, how is that not progress?
Confidence scams have always been a problem even for other Nigerians. We tend to be very suspicious of people on average.
A couple of years back my father, who teaches at a Nigerian university, got a call from his head of department who was in Europe at the time. He needed help clearing some important items that were being held by customs who were demanding payment to release a moderately valuable shipment into the country while he was away.
My dad was put in contact with an agent who requested some money to sort things out. At first, it was a token amount, but things escalated gradually to the point where he was considering "borrowing" money from both my sister and myself (we were both on scholarships while at university). My sister was home at the time and grew suspicious from the outset but my dad was totally convinced that the voice was authentic and that the caller simply knew too much. It was actually offensive for my sister to presume that he was that naive.
Fortunately, he fought the need to save face by not disappointing his boss in a crisis and told the agent that he couldn't send any more money. The numbers were disconnected immediately afterwards and his Head of department had no idea what he was talking about when he eventually returned. He didn't push the matter any further as it would be embarrassing to admit that he was conned.
This was a well calibrated psychological attack designed to take advantage of his personality and reputation as a man that can be relied on. It also preyed on his need to maintain his social standing among his peers. He had nothing to gain from this other than gratitude and an appeal to greed would have set off red flags immediately.
My conclusion was that this was probably engineered by some students who knew them both fairly well.
419 scams are fairly common in Nigeria and you become a target if you are perceived to be wealthy or recently come into some money. The "Yahoo boys" are ridiculous though because they are fairly easy to spot and are only marginally literate. It is part of a foreign condescending attitude to assume that that level of poor presentation is an acceptable standard for an educated Nigerian. Most of us think people who are scammed are suckers but it's an evolutionary game and some of the scammers are really efficient predators.
* the 75 most common words make up 40% of occurrences
* the 200 most common words make up 50% of occurrences
* the 524 most common words make up 60% of occurrences
* the 1257 most common words make up 70% of occurrences
* the 2925 most common words make up 80% of occurrences
Speculatively, similar frequency analyses of library calls in various codebases for a programming language may reveal a similar long-tail distribution. This would provide a list of high frequency library routines that would be worth memorising - with clearly diminishing returns as the vocabulary grows.This kind of analysis would also point at important targets for optimisation, simplification and parallelisation as well as direct the efforts of library implementers for new languages.
The whole thing is a vicious circle, this kind of news has to be sensationalised because we have been steadily desensitised by the media.
As a teenager in the 90s, I remember getting mildly claustrophobic at the thought of watching another Movie or TV Show about a hijacked plane or hostage situation. There were just so many building up to 9/11. Essentially, the messages were things like 'Enemies of America hijack planes', 'Radical citizens plant bombs', and now 'Angry employees/students show up with automatic weapons and body armour.' These are as much statements as they are calls to action.
Let's take a look at this year's Emmy nominations for the best actor category, these actors play: a ruthless corrupt bootlegging politician-gangster, a high school chemistry teacher turned murdering drug kingpin, a vigilante serial killer working in law enforcement, a terrorist war-hero congressman secretly radicalised in captivity and an alcoholic misogynistic deserter ad executive with a stolen identity. The only leading character that does not belong in prison is from a British production. Perhaps this category should be renamed the 'Best Magnificent Bastard' award.
America has this fascination with revealing its weaknesses to 'enemies - both foreign and domestic' and celebrating monsters. Most people learn early on that, unless one is trying to play and remain the victim, painful memories and weaknesses should not be put on public display for the casual friend, enemy and stranger.
As far as I know there is no Vala runtime or garbage collector. Memory management is by automatic reference counting but there is support for manual C++-style pointer syntax with new and delete operators.
I agree that Vala was not specifically designed for 'low-level systems stuff' but I see no reason why this is not technically possible. The Vala compiler is a source-to-source translator to C and in my previous experience worked fairly well interfacing with fairly low-level C libraries like MPI and OpenCL. I assume that a kernel writer does not require much more than a limited subset of the language that supports C pointers, structs and arrays.
For transparent access to compiled C libraries, external C functions need to be declared in a 'vapi' file and although these were already provided for a large part of the GNOME ecosystem, documentation on vapi syntax was nearly non-existent. Object-orientation and standard libraries are integrated with Gnome's GObject system and Glib libraries which I found unfamiliar and usually better documented in C at the time.
GPU computing is promising, but given that it is fairly difficult to predict how much of a speed-up to expect before the actual work is done, I find myself asking whether some of the expended effort is really worth the trouble. It seems that there are two classes of applications where this makes sense:
1) Minimising latency/increasing responsiveness for smaller algorithms such as in a user application or service.
2) Doing large volumes of computation in a high-throughput system.
At the moment, the CUDA platform is well ahead of OpenCL in terms of maturity, features, tools and documentation. However, OpenCL runs on CPUs, both NVIDIA and AMD GPUs and work is being done towards targeting FPGAs [1]. Interestingly, Clang also supports compiling OpenCL kernels directly to native code.
In all likelihood these platforms will stay outside the mainstream until better abstraction layers exist to shield programmers from the low-level architectural details without sacrificing performance. Something similar to the directive-based OpenACC standard that is able to perform close to hand-optimised close would go a long way.
I only "speak" CUDA, however the primitives provided by the CUDA/OpenCL development platforms are largely equivalent. For a single "kernel" the GPU provides parallelism at two levels,
1. Fine-grained Thread-level parallelism: You have a group of threads performing ideally identical operations. The model is called SIMT (Single Instruction Multiple Thread) because it is not quite SIMD but close. Say you have (1-1024) threads, the hardware splits these into warps of 32 threads. The 32 threads within a warp execute together in classic SIMD manner, however only one of these warps is executing on the actual hardware at any given time. When a warp makes a request for data from high-latency device memory, it is typically suspended and another warp swapped in. This allows latency hiding between warps. However, calling __syncthreads() forces all warps to synchronise. Any conditional statements cause warp divergence, the warp is split and the alternative execution paths run sequentially. Clearly, this incurs a performance penalty as SIMD behaviour is lost and should be avoided as much as possible. The 'ideal' identical operation size is therefore a warp or 32 consecutive threads, however occasional divergence is acceptable. Memory access patterns are much more important however.
2. Coarse-grained Block-level parallelism. A block is an independent group of threads. The platform allows mutiple blocks to execute on different Streaming Multiprocessors. The hardware may schedule multiple blocks to execute on a single SM if there are sufficient resources (such as register and shared memory), this also facilitates latency hiding. Other than the kernel invocation which created multiple threads blocks, they are unrelated and cannot communicate or be synchronised.
The potential thread level parallelism is what should be considered. If your algorithm is representable as fine-grained floating point vector operations without excessive divergence then it is well-suited to the GPU. Of course, the actual performance gain depends on the quality of implementation and optimisation is highly GPU-specific and non-trivial. I recommend
http://developer.download.nvidia.com/compute/DevZone/C/html/...
as an example of how important optimisations lead to those impressive performance numbers.
Also, content-addressability may be implemented fairly independently of the actual structure, that is an encoding problem that is analogous to an error correcting code in a higher dimensional vector space.