Hacker Newsnew | past | comments | ask | show | jobs | submit | more ppoint's commentslogin

Are E5 Skylakes available exclusively to Google? Can't find E5 v5 on Intel's site: http://ark.intel.com/#@Processors.


Purley is the big Skylake based server platform chipset reset for Xeon E5 and higher so Google must have early access to it.

Skylake Xeon E3 is on the current older server chipsets..

Purley is supposed to be a significant improvement.


We announced with Intel back in November that we would be getting early access to Skylake [1]. Today's announcement is that we're ready to let more people in the door.

[1] https://cloudplatform.googleblog.com/2016/11/power-up-your-G...


Thanks!


E3 is not a server SKU, it's meant for workstations and it's more similar to i7 than E5. It lacks various features including ECC RAM which, I suppose, is a no-starter for cloud hosting. Also some virtualization capabilities are missing, notably accelerated interrupt injection (APICv+posted interrupts).


E3 does support ECC RAM. I'm running a Haswell Xeon E3-1225 with 16GB ECC RAM as my workstation.


I believe that the ability to use ECC RAM is actually the main differentiator between i7 and E3.


I've seen E3's in low end servers like the Dell T20...


I've only seen E3s on Intel's web-site. But maybe E5s are coming? That'd be awesome.


I think it's relying on libcrypto from openssl, at least that was the case when it came out. E.g. it implements TLS/SSL protocols only, no crypto. Not sure if that's changed and they implemented crypto as well.


From a brief perusal of the source code[0], it still does. It makes this claim on the `README.md` seem rather disingenuous:

> One of the key benefits to s2n is far less code surface, with approximately 6,000 lines of code (compared to OpenSSL’s approximately 500,000 lines).

0: https://github.com/awslabs/s2n/blob/master/crypto/s2n_dhe.h#...


calloc zeroes memory on allocation.


Yes, I think the question was something like "why doesn't malloc call calloc?".


Always nice to have options. Not zeroing memory on allocation might save a few cpu cycles.


It's pretty much the definition of false economy. Would you rather save a few cycles or suffer debilitating security bugs at random intervals? Always use calloc unless a) there's a proven performance problem and b) you know for a fact that due to careful inspection/static analysis/black magic malloc is safe. Then use calloc anyway because why risk it?


It depends on the size of the chunk of allocated memory. If it is quite large, time spent zeroing it can be substantial. Then again, if you're allocating in performance critical path, you're doing it wrong anyways.


Tempered Networks | Senior Software Engineer, data plane networking | Seattle, WA | onsite, full time, https://www.temperednetworks.com

Looking for experienced networking data plane software engineer: Linux kernel networking stack, userspace networking, crypto.

Official position description here: http://www.temperednetworks.com/senior-swe-dataplane/

Send an email with your resume to jobs@temperednetworks.com (make sure to mention HN in the email).


Are you planning on adding code search feature any time soon (in the cloud, that is)?


Bitbucket PM here, We are currently working on code search for Bitbucket cloud, you can follow our progress here: https://bitbucket.org/site/master/issues/2874/ability-to-sea...


> The DOL approves LCA based on minimum wages

Is this true? I thought that the salary of H1b worker must match an average paid for this kind of position in this area. In fact, I'm fairly certain that the company sponsoring H1B must supply documents showing they did the research, on both wages and the fact that they couldn't find anyone in the US to do this kind of work.


The wage for a Level 1 'Computer Programmer' in SF is 66k/yr[0].

The wage for a Level 4 'Software Developers, Applications' in SF is 138k/yr[1]

Which option do you think employers choose?

[0] http://www.flcdatacenter.com/OesQuickResults.aspx?code=15-11...

[1] http://www.flcdatacenter.com/OesQuickResults.aspx?code=15-11...


Ah, I guess if you do want to game the system there is plenty of room to do so.


From what I've seen at my employer, H1Bs end up getting paid less simply due having a different set of priorities. We don't discriminate between citizens and H1Bs when crafting an offer for employment, but citizens are far more likely to try to negotiate for more compensation whereas H1Bs are more interested in getting an EB2 and are even willing to sacrifice salary to get that.

It wasn't until I hired a few H1Bs that I understood that the H1B discount isn't just a simple wrong. When looking at H1B compensation, we should probably also include additional legal fees paid on behalf of the worker rather than simply comparing salary to salary if we want an apples-to-apples comparison.



Great stuff! Would be great if similar comparison existed for data grids (tables).



How does this compare to DPDK?


this is seems to be an offshoot/extension of pf-ring.

apart from performing packet operations in userland also defines a language with functional-composition (take a packet, return an 'enriched' packet) for further processing e.g. fanout/io etc.

ast from pfq/lang application gets translated into appropriate structures, in the kernel, and gets executed on top of the driver...


Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: