Transmeta Crusoe, it would be very useful today for running Docker containers built for different instruction sets without being as slow as QEMU. The performance penalty was only about 20% and the power efficiency was actually better than native x86.
I'm guessing he means a "Real-time Operational Database". This seems to be a generic term for system like a data warehouse that contains current data, instead of just historical data. If you are taking the output of a Spark flow and storing it in Postgres or MongoDB or HBase for applications to query, then those could be considered RODBs.
Since this is Amazon, I suspect he is referring to SPICE (or their internal version), which was released last fall as part of AWS's QuickSight BI offering...
"SPICE: One of the key ingredients that make QuickSight so powerful is the Super-fast, Parallel, In-memory Calculation Engine (SPICE). SPICE is a new technology built from the ground up by the same team that has also built technologies such as DynamoDB, Amazon Redshift, and Amazon Aurora. SPICE enables QuickSight to scale to many terabytes of analytical data and deliver response time for most visualization queries in milliseconds. When you point QuickSight to a data source, data is automatically ingested into SPICE for optimal analytical query performance. SPICE uses a combination of columnar storage, in-memory technologies enabled through the latest hardware innovations, machine code generation, and data compression to allow users to run interactive queries on large datasets and get rapid responses."
This is a gnarly run. Almost 7000 ft of climbing (he gives the value in m, but all my reference points are in feet). It appears he was not auto-pausing, but did stop quite a bit. The time spent moving was 5:49. His pace, excluding the time stopped, was 11:12 min/mi. With the elevation gain, this is most definitely running.
For reference on how to normalize the elevation gain, I like to add 1.5 miles for every 1000 feet of elevation gain/loss (combo). So, in the way I plan, I'd look at that 50k like a 41 mile run on flat ground.
I can't comment for running, but for backpacking we would add a mile for every 1000 feet of gain, ignoring loss. So 500 up, 400 down, 500 up feels like an extra mile. You have to look at elevation maps to get the nitty-gritty, but I've found it to be pretty fair.
To a decent runner that is nearly walking though. A record road race marathon pace is ~4:45 min miles. A slightly above average runner can do say 8 min miles over 26.2 miles on flats, which is only a titch under the 50k of this race.
Trail races are slower, and 7000 feet is a decent amount of climbing.. but he was in no danger of setting any land speed record. 15 min miles is not that much faster than walking.
So this proves you can speed-walk an ultra in a fasted state. Which is cool and shows how strong our bodies are. scott jurek is amazing at ultras, and a vegan - but he definately eats and runs off sugar.
> A slightly above average runner can do say 8 min miles over 26.2 miles on flats
What is it about the running community that makes them say ridiculous things like this?
The fastest average marathon pace I could find anywhere was 9m 6s for 20-year old men.[0] To say that it takes only someone "slightly" above average to maintain a pace 12% faster than that for an entire marathon is ridiculous. Just because a lot of people do it every year doesn't mean that it's a great accomplishment. For context, that means someone finished nearly half an hour sooner.
3200 people there finished. So let's take middle, #1600: 4:11, which is 9:34 pace.
8 min pace is 3:29, which is something like finish 700 / 3200 in that age group.
So to me, finishing in 700th place out of 3400 people in a race is "slightly above average". But who cares, even the 9:30 pace is far above the 16 minute pace of OP.
Realize too, that many people running a marathon are running their 1 lifetime marathon - they never ever plan to run another. Drop all those out, and you will find 3:30 may be about average or even a tad worse among people who are serious runners.
I agree, my point was that the fastest "average" time I could find (regardless of quality of the source) was still sufficiently slow enough compared to the 8:00 pace figure that 8:00 is in fact not slightly above average.
Just as a point of reference I ran this 30k [1] (18.6 miles) at 8:38 pace vs. a flat road half marathon (13.1 miles) at 6:41 pace about 5 months later.
As you can see in this [2] it climbs from a low of 2500 feet to 3700 feet. I'd estimate 1500 feet total so much less than in the article.
I usually place in the top 5% of men in most races.
I'm not a huge runner, but an 11 minute mile over that distance is great. An 8 min/mile marathon run is a highly experienced runner and not just someone slightly above average.
> A slightly above average runner can do say 8 min miles over 26.2 miles on flats
I make that a 3:30 marathon which I would suggest is rather more than "slightly above average" - you'd be in the first 25% male finishers at London 2015, for example.
I consider 25th - 40th percentiles as slightly above average... as many many of those people in your race are running their first and only marathon. Of people who do at least 1 marathon a year, I don't think you will find a 3:30 marathon much past the middlepoint. Not even close to an open BQ time for example.
While your intuition here may well be valid, to an outside observer it looks like you're cherry picking numbers without justification. Why 25th-40th percentiles? Do we know the distribution of first-and-only marathon runners to repeat runners or is this a pure intuition? If the latter I would be concerned about the flaws that exist in human thinking when it comes to generalizing over groups.
Why not? Someone wanted to quantify "slightly faster than average", so I made up some percentiles. I don't think there is a scientific definition of "slightly faster than average" so I think I can do this. You could I suppose counter with you think it means "35th-49th percentiles" which is valid but doesn't really change the argument too much.
> Do we know the distribution of first-and-only marathon runners to repeat runners or is this a pure intuition?
Just intuition. It would be a tricky thing to survey, as you would need to wait for all current people who have ran exactly 1 marathon to die, to confirm they do not indeed run more .
> Do we know the distribution of first-and-only marathon runners to repeat runners
I did look for those stats for London but I couldn't find anything relevant - I'd assume they collect that information on the entry form but it's possible they don't or just don't care to publicise it later.
> The average finishing time globally for 26.2 miles in 2014 was 4hr 21min 21sec – about 40 seconds faster than the average for the period 2009-2013. Men’s average finishing time was 4hr 13min 23sec, while women’s was 4h 42min 33sec – 29min 10sec slower.
I'd suggest your 3:30 was still definitely more than "slightly above average".
If you want to quibble about the median instead of the mean,
pneumatics corrects me and says his actual moving pace was 11:12 min/mi. This is a pretty good pace, given the terrain, and far beyond speed walking.
I've run everything from flat, paved marathons to alpine mountain ultras. The paces are totally different. I'm going to be running a technical ultra a good 3-4 minutes/min slower than I'd run a flat road marathon.
The owner must be present during home-sharing, so each property owner should be limited to sharing one property at a time. Since the properties must be licensed, it should be easy to shut down the multiple-property Airbnb landlords.
What this might enable is "sharing" a huge property, say an apartment complex, essentially opening up a new class of low- or no-amenity hotels.
| I don't know a good introductory book, unfortunately.
I liked Eric Schechter's Classical and Non-Classical Logics for an eye-opening view into how logical systems are constructed from axioms. Might be a satisfying read for OP, as well.
This is an important lesson, but I was hoping for an implementation of stopping rules. If you're looking to actually stop your N-armed bandit trials properly, look at the S+ package SeqTrial. It implements the stopping rule framework described in KittelsonEmerson99 (http://www.ncbi.nlm.nih.gov/pubmed/11315020).
I installed a full application to do this the last time I needed annotated screenshots that didn't look like they were made by a child. I used Jing fwiw.
