Monthly Archives: November 2012

heliod web server: fast then, still fast today

Quick post – as an update to my post here, Jyri Virkki has published a comprehensive set of benchmarks, comparing heliod’s out-of-the-box performance to all the other current popular HTTP servers out there. Considering the last public comparison I could find of what was then Sun Java System Web server vs Apache was in 2007, these new results are highly interesting:

“heliod had the highest throughput at every point tested in these runs. It is slightly faster than nginx at sequential requests (one client) and then pulls away.

“heliod is also quite efficient in CPU consumption. Up to four concurrent clients it is the lightest user of CPU cycles even though it produced higher throughput than all the others. At higher concurrencies, it used slightly more CPU than nginx/lighttpd although it makes up for it with far higher throughput.

“heliod was also the only server able to saturate the gigabit connection (at over 97% utilization). Given that there is 62% idle CPU left at that point, I suspect if I had more bandwidth heliod might be able to score even higher on this machine.

“These results should not be much of a surprise… after all heliod is not new, it is the same code that has been setting benchmark records for over ten years (it just wasn’t open source back then). Fast then, still fast today.

You can read the total run of tests plus information graphs at Jyri’s blog entry: http://173.255.252.27/jyri/articles/.

 

Incidentally, I came across a blog post from someone who was also apparently on the Sun Java System Web Server group at Sun, who states:

“Since Oracle no longer offers updates to individual users, and refuses to respond to requests for information about how an individual can acquire the updates, I have elected to stop writing about the server. If the moribund Open Web Server gets branched I will happily contribute to the pool of knowledge that exists for it.”

Hmm, maybe someone should give him a heads-up about heliod…

Advertisements

System hard freezes with the AMD FX-8350

As an update to my post here, I observed seemingly random freezes on my system upgraded with the AMD FX-8350. The behaviour encountered was a total freeze of the desktop environment, no response to local keyboard nor mouse, no response to attempting to launch a virtual console, no reponse to pings over the network, and no ability to log in remotely. The only way to restore system operation was to perform a hard reset. Interestingly I could also consistently crash the system running a GraphicsMagick benchmark. Additionally, the freezes were OS-agnostic, occurring under both OpenIndiana and Ubuntu Linux.

Looking around online you can find several posts from folks on AMD Bulldozer rigs with very similar issues (such as detailed here), including a few from people who have rather alarmingly downgraded to a Phenom or Intel CPU as a “fix”, after having received advice to alternately update the motherboard BIOS, faff around with multiple BIOS settings, test and replace the RAM, power supply and hard disk, RMA-ing the new CPU (!?), and on and on and on. Most of this didn’t really add up, and similarly my problems were encountered on a system that was hitherto generally stable using an older-generation CPU (the Phenom II X6 in my case).

To cut a long story short, this quite simply turned out to be the motherboard not stably supporting the FX-8350. Although the ASRock 870iCafe 2.0 is an AM3+ compatible part and advertised as being “8 Core Ready” (to the point of specifically claiming compatibility with the FX-8350), the reality is that the latest BIOS release was in December of 2011 – a major red flag. After upgrading my motherboard to a Gigabyte GA-990FXA-UD3 with the recent F9 BIOS, the system is now stable. And yes, this is using the original PSU, RAM, graphics card etc.

For the OpenIndiana readers, the GA-990FXA-UD3 works fine, although don’t expect USB3.0 support:

Gigabyte GA-990FX-UD3 driver support on OpenIndiana