Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Google Cloud Platform is the first cloud provider to offer Intel Skylake

cloudplatform.googleblog.com

Post reply on HN
Print view View original

151–160 of 214 posts

Re: Google Cloud Platform is the first cloud provider to offer Intel Skylake

#151

Post by jdhawk »

eloff wrote: post #126
zbjornson wrote: post #18

I've been benchmarking these against Haswell and Broadwells. Despite being 300 MHz slower, we're getting between 5 and 45% faster benchmarks on linear algebra functions that we run a lot, even without doing much work to tailor to AVX512 instructions yet. The cache is also a whopping 56 MB.

Still limited to a measly 208gb of RAM? That was pretty good back in the day, but now my desktop workstation has 128gb. AWS offers up to 488gb on their latest R4 instances, and 2tb on their X1 instances released a year ago.

No - I believe these will be up to 64core & 416GB

Re: Google Cloud Platform is the first cloud provider to offer Intel Skylake

#152

Post by doh »

nugator wrote: post #144
doh wrote: post #84

Earlier quoted context omitted.

Spending on GCP is more predictable than on AWS. Also the general behavior is more predictable. For instance all pre-emptible instances are terminated in 24 hours. On AWS based on when they need the resources. I believe that whatever GCP does, does better. Faster network, faster spinning and turning off servers, simpler quotas, etc. GCP doesn't offer that many services as AWS and most of those are not very compatible…

Holy cow, 28 000 servers, can you give us a clue to what you need so many servers for?

We're processing videos, a lot of them. Roughly 34PB worth of data every month.

Additionally we're running our search engine at top of it with couple of other services.

You can check what we're up to here https://pex.com

Re: Google Cloud Platform is the first cloud provider to offer Intel Skylake

#153

Post by petra »

reitzensteinm wrote: post #24
lightedman wrote: post #17

I'd rather wait for Ryzen. You won't know which Skylake processor you're getting - the gimped one or the non-gimped one. AMD tends to keep their features consistent across the line.

Depends on the workload. A compute bound vectorizable problem could be four times faster on Skylake.

Why 4 and not 2 (512/256)?

Re: Google Cloud Platform is the first cloud provider to offer Intel Skylake

#154

Post by nugator »

doh wrote: post #152
nugator wrote: post #144

Earlier quoted context omitted.

Holy cow, 28 000 servers, can you give us a clue to what you need so many servers for?

We're processing videos, a lot of them. Roughly 34PB worth of data every month. Additionally we're running our search engine at top of it with couple of other services. You can check what we're up to here https://pex.com

Cool, thanks for the info!

Re: Google Cloud Platform is the first cloud provider to offer Intel Skylake

#155

Post by boulos »

gkop wrote: post #102
zbjornson wrote: post #94

Earlier quoted context omitted.

As far as HIPAA compliance goes, though, AWS EC2 requires dedicated tenancy, whereas GCE allows HIPAA workloads on any VM, including preemptible VMs.

This is true, however Amazon's standard BAA is a bit more robust than Google's. This didn't stop my HIPAA-compliant company from choosing Google, but it does make it more difficult to compare the two clouds on HIPAA-friendliness.

Can you say more about what you found less robust? And (hopefully) were you unable to get the legal agreement up to snuff by the end?

Disclosure: I work on Google Cloud (but not compliance).

Re: Google Cloud Platform is the first cloud provider to offer Intel Skylake

#156

Post by SomeStupidPoint »

boulos wrote: post #119

There are a few threads asking about various features (SGX, TSX, etc.) so I want to make a top-level comment: we're not ready to share more today (sorry). Disclosure: I work on Google Cloud.

It's okay you're not ready to share more today, but do you know/have an idea of when you will be?

Re: Google Cloud Platform is the first cloud provider to offer Intel Skylake

#157

Post by eblanshey »

briankereszturi wrote: post #52
rattray wrote: post #20

Earlier quoted context omitted.

Out of curiosity, why would you prefer Google Cloud to AWS?

Managed Kubernetes.

What's the difference between using Kubernetes on GCP and running Kubernetes on a node on AWS EC2? By "managed" do you just mean I don't need to manage the lead node?

Just getting into Kubernetes :)

Re: Google Cloud Platform is the first cloud provider to offer Intel Skylake

#158

Post by briankereszturi »

eblanshey wrote: post #157
briankereszturi wrote: post #52

Earlier quoted context omitted.

Managed Kubernetes.

What's the difference between using Kubernetes on GCP and running Kubernetes on a node on AWS EC2? By "managed" do you just mean I don't need to manage the lead node? Just getting into Kubernetes :)

That's correct. The master nodes are managed by GCP and upgrades to the kubernetes system are trivial.

Re: Google Cloud Platform is the first cloud provider to offer Intel Skylake

#159

Post by wlamond »

boulos wrote: post #132
ajaimk wrote: post #130

Earlier quoted context omitted.

Bothers me a bit with "32 cores" listed when they are 16 core/32 Thread machines.

I try to consistently say "vCPUs", but yeah...

Given the marketing toward HPC tasks, shouldn't hyper-threading be disabled?

Re: Google Cloud Platform is the first cloud provider to offer Intel Skylake

#160

Post by boulos »

wlamond wrote: post #159
boulos wrote: post #132

Earlier quoted context omitted.

I try to consistently say "vCPUs", but yeah...

Given the marketing toward HPC tasks, shouldn't hyper-threading be disabled?

No? We make sure that pairs of vCPUs (hypertwins) stay together. If you'd like, you can offline your threads in the guest for a fairly similar outcome. Some of us regret that we sell an n1-standard-1 and the shared core machine types, but it's pretty nice to have a $5/month VM price point for those that want it. Most folks doing serious numerical computing end up using much larger VMs, and always crossing NUMA nodes (which is a much larger impact).

Disclosure: I work on Google Cloud (and historically focused solely on Compute Engine).

Post reply on HN