Try new Power8 CPU for free on RunAbove cloud: 176 threads! | Hacker News Reader

Try new Power8 CPU for free on RunAbove cloud: 176 threads!

labs.runabove.com

58 pointsbaqs28 comments

Open articleView on HN

Quick benchmark, compared to this one Intel Xeon E5-1620v2 4c/8t 3,7 GHz+/3,9 GHz+ [0]

#Intel

sysbench --test=cpu --cpu-max-prime=100000 --num-threads=8 run execution time (avg/stddev): 29.8777/0.01

#Power8

sysbench --test=cpu --cpu-max-prime=100000 run --num-threads=8 execution time (avg/stddev): 9.5050/0.00

sysbench --test=cpu --cpu-max-prime=100000 run --num-threads=16 execution time (avg/stddev): 4.7733/0.01

sysbench --test=cpu --cpu-max-prime=100000 run --num-threads=126 execution time (avg/stddev): 2.6502/0.14

[0] http://www.ovh.com/us/dedicated-servers/infra/2014-EG-32.xml

Pretty impressive

I wonder what the temps look like and how it would perform at 3D rendering and video encoding.

I would be really interested for a database server, right now they offer a single ssd (no RAID), this is a no-go for a production db.

Not exactly, it says "Local RAID" on the Power8 lab page. So it's a RAID of HDD.

yes, I forgot it is actually a single virtual machine running on a physical server.

The main disk is /dev/vda2

They run hot. Yes, they've been improving on heat output and power consumption, but they're certainly not anywhere near most Intel parts — that said, they're chasing a market where perf matters above all else.

Video encoding probably wouldn't be so great comparatively, in large part because I don't think any encoders use the POWER8 vector stuff.

Wow, that seems almost unbelievable. Power 8 is able to beat Intel architecture by a factor of 0.3 at the same level of concurrency? Can someone offer an explanation of how/why this happens, but we're not all running Power architecture on our servers?

One reason would be that power uses 8 way smt while Intel (with hyper threading) uses 2 way smt. If your code has tons of cache misses you will gain greatly from 8way smt since another thread can run on the physical core while other threads are waiting for memory. I suspect PPC would do very poorly with High performance code with low cache misses. However for something like a Java program where cache misses are abundant I would expect to see a big boost in performance. The benchmarks above are therefore completely dependant on the cache size.

I worked with the people at IBM who did optimization work for Java on Power. The ridiculous amount of work they did was amazing. There's dozens upon dozens of compiler papers and presentations that are more or less power specific. It was an exciting, and stressful, thing to be a part of.

I think if you look closely at the specs, I think the Xeon has about 2.5MB of L3 per core, while P8 has about 8MB of L3/core. There are bigger differences than this. But the examples above have both processors set at SMT2 (~30 seconds vs ~10 seconds). Differences are greater for SMT4 and SMT8. Of course, if your code fits nicely into the L3 - e.g. if your hot code consumes 3MB, then Power8 will be a bigger winner since there may be more thrashing for the x86 chip - unless, of course, the prediction routines is straight forward and the cacheline can be prefetched (think sequential access vs random access).

From what I saw elsewhere, highly parallel code runs faster on P8, where as single thread perf is faster on x86. So if your HPC app is basically single threaded on one of your compute nodes, then that would be faster. But if your HPC code is highly multithreaded on the core, then P8 may surprise you.

Here is the /proc/cpuinfo:

    processor	: 0
    cpu		: POWER8E (raw), altivec supported
    clock		: 3026.000000MHz
    revision	: 2.1 (pvr 004b 0201)

    timebase	: 512000000
    platform	: pSeries
    model	: IBM pSeries (emulated by qemu)
    machine	: CHRP IBM pSeries (emulated by qemu)

The OS (Fedora 19) runs on Big-Endian ABI.

Python is available out-of-box.

Ruby, Java (OpenJDK) and Nginx are available from yum, Node.js and Mono are not.

Make sure to provision a huge screen if you ever plan to run htop. Or you might not even be able to see the top of the process list...

Or just Setup -> Meters -> CPUs bar -> Remove (However leave the CPU (average) bar)

Can anybody explain the sorts of workloads for which the Power8 architecture would be a cost-effective option?

It's useful for hightly intensive workload that can't be distributed but that can use multicore. Like SQL databases, video encoding, etc...

Can you run Java with full JIT on Power8? What about nodejs? I know IBM did some talks/slideshares of porting nodejs to power.

Yes, Java and Node work.

I've just read their main website. Out bandwidth at $0.01 per GB? Has anybody tried this? Looks too good to be true.

Disclamer : working @ovh

> "Looks too good to be true"

It's not :) we built a brand new network to be able to offer this at scale.

Brilliant, I can finally implement one of my bandwidth intensive ideas! Thanks Steven!

You're welcome ! feel free to feedback your impressions or remarks @runabove : https://twitter.com/runabove

Is this "brand new" network substantively different from the one you launched last year? Looks to me like this is still based off of OVH's (oft never fully online) internal 1Tbps network + a lot of rented fiber or virtual-PoP over rented transit.

Is the bandwidth still part of the lackluster Volume network you use in your French DCs?

Absolutely massive!

Well, here is what htop looks like with 176 threads: https://community.runabove.com/kb/images/2014-10-06-power8/h...

The creators of htop just need to get inventive on how to present the data. One idea is to change it into a histogram of percentages (these many threads @ this level of use).

RunAbove recently released a $2.50/month VPS, but they have also a power8 offer. This seems like an exclusivity.

Anybody tried it already ?

It's just another shell company for OVH, not an exclusive they just buy a lot of CPUs (especially Atom chips)

Also, it seems like you and a few others are sockpuppet accounts

FYI I am currently working there and can give you some proper benchmark results running MySQL on power8: https://twitter.com/o_erreip/status/521411270411816962

We are trying to optimize it with Stewart Smith, who achieved 1 million req/s on p8

Reply on news.ycombinator.com