2022 Cloud Report(cockroachlabs.com) |
2022 Cloud Report(cockroachlabs.com) |
Here is the direct link to the report
https://www.cockroachlabs.com/pdf/2022-cloud-report-cockroac...
Don't underestimate the damage an overly eager and naive developer can do.
Or, (maybe more likely in this case?) a CMS author who's been given a little bit too much power?
document.querySelectorAll('.aos-init').forEach(e => e.remove())
The above will remove all of those popups.
> Overall, the gap for most AMD-based processors closed almost immediately when we controlled for NUMA nodes – in other words, when we only considered runs where each instance showed all vCPUs running across a single NUMA node. When we did this, the performance gap dropped from 22% to 1%, which is smaller than our margin of error.
How does one avoid machines with vCPUs across multiple NUMA nodes? Do you just spin the machine up, run `lscpu | grep 'NUMA node(s)'` and kill the machine if the value reported is anything but 1 and try to spin a VM again?
I would love to see an AMD chip as fully integrated as an M1, moving the RAM fully on die and part of the Infinity fabric directly. The insane memory bandwidth of the M1 is what keeps it competitive.
You are trying to “make” the users click “Allow all” just to hide this trash as quickly as possible. It is low.
>We chose not to test ARM instance types this year as CockroachDB still does not provide official binaries for that processor platform. Official support for ARM binaries is slated for our Fall release (22.2), so we expect to return to testing this processor platform next year.
It's not that easy to displace x86...
https://www.phoronix.com/scan.php?page=article&item=graviton...
Amazon here seems to be leading the Arm technology pack, as they are using a newer generation of CPU, while arm instances on other providers tend to be providing gravaton2 (Neoverse-N1) based instances. I would imagine that gradually changing in the near future as those vendors also upgrade.
Honeycomb has some great blogs about it: https://www.honeycomb.io/blog/present-future-arm-aws-gravito...
(If I were in the market for their database and the alternatives were close in parity and price, I'd be likely to choose a competitor for the name not being "cockroach" alone.)
Partitioning for existing DBs: Citus for Postgres and Vitess for MySQL.
Snowflake apparently just started to target transactional workloads with UniStore.
Sometimes we get lost in the criticism of every little thing that these companies do and forget that honestly, they're all cranking out great products.
I know this has been said a million times, but it's worth repeating because somehow the idea is still floating around – the M series very much does not have the RAM on-die. It's not even in the same package – it's standard LPDDR4/5 sitting off to the side with a lot of channels.
[1]: https://www.bgr.in/top-products/best-phones-with-lpddr5-ram-...
[2]: See other thread. :)
Current rumours suggest that's where AMD is heading, Zen 5 having multiple accelerators integrated and Zen 6 having HBM part of the package (on the datacenter variants):
I am not a fanboy, but a realistic dude.
AMD ruled the last years but Alder Lake overtook Vermeer on both performance and price/performance.
And that is with a process node difference, Intel using 10nm vs AMD using 7nm.
And the future looks like Intel will enhance the distance between its performance and AMD's.
Alder Lake relies on brute force, inefficient power consumption to regain the performance crown. AMD's chips are much more efficient, and efficiency matters in datacenters. There is only so much power and cooling available to each rack unit.
I think Sapphire Rapids holds a lot of promise, but it remains to be seen.
AWS revenue is 62B. Azure Cloud revenue is 23B. GCS revenue is 22B. I'd say that at 42% of the market, other cloud provider service offerings cumulatively are used by a significant amount of enterprises as well. These revenue numbers are probably artificially deflated due to the land war going on between these business units (i.e. Azure/GCS have to offer larger discounts to steal business from AWS) so it's hard to actually compare without deeper analytics access into these providers. Regardless, even individually these are massive revenue numbers indicating there's plenty of deployments on each of the clouds. Also, the "nobody got fired for IBM" reasoning is pretty flawed in tech; it's kind of ridiculous how often this argument is made.
Just to preface this next bit. I'm not asking about scenarios like "I run all parts of my business on AWS" which could change the benefits towards AWS (vendors love lockin because it raises prices). I'm asking mainly about a greenfield project that is selecting between the clouds equally.
> Each cloud comes with unique service / API complexity and despite being managed services that experience does not translate 1:1 across clouds
AFAIK the vast majority of all of this has been thoroughly commoditized. Do you have anything specific that you see as an important product/feature that Amazon has that the other's don't?
> AWS IAM policies cannot be reused
Not relevant for new deployments & AWS IAM policies are famously overly complex resulting in security vulnerabilities due to misconfiguration, but sure. IAM policies are unique to each cloud provider.
> and there may be differences in availability, durability, feature set et al
Experientially it seems like Google has an advantage there - AWS seems to have more frequent and longer-lasting outages. I doubt durability has a major difference & feature sets tend to be fairly even between them (might matter for the long tail). Also, that statement feels kind of empty as all it says is that there "may be" differences without stating any or which provider might come out ahead.
ARM isn't the optimal solution for every application at this time, but anyone who isn't seriously considering it probably needs to update their information.
Intel didn't use EUV, so no.
[0]: https://drafts.csswg.org/mediaqueries-5/#prefers-reduced-mot...
[1]: https://vestibular.org/article/what-is-vestibular/about-vest...
Or just follow the spec.
> In most cases, it just blocks or hides cookie related pop-ups. When it's needed for the website to work properly, it will automatically accept the cookie policy for you (sometimes it will accept all and sometimes only necessary cookie categories, depending on what's easier to do). It doesn't delete cookies.
It would be helpful to have a "what did you choose on this site" view though if that doesn't exist yet.
And in desktops performance is what matters. And price/performance ratio, and both are in Intel's favor.
Both Intel and AMD have plans to integrate memory more tightly onto the package of their datacenter processors in the next couple of years, IIRC, and that seems to be what the OP of this comment thread was hoping they would learn from M1.
But, whatever.
https://github.com/cockroachdb/cockroach/issues/62903#issuec...
Monopolies don't deserve credit.
The idea that ASML would somehow be more worthy of merit if Nikon/Canon also succeeded is weird.
Would the moon landing me more impressive if Russia and Japan also managed it?
At the height of the craze, I saw a five-year-old card, the GTX 1050 (not even Ti!), going for over 200.
https://pcpartpicker.com/trends/price/video-card/
From February: https://www.techradar.com/deals/looking-for-a-gpu-dont-buy-y...
Consumer PCs don't match this bandwidth because DDR DIMMs generally aren't as fast as LPDDR. Plus AMD & Intel limit their mainstream consumer CPUs to two memory channels, both for cost savings and to segment the market.
What about LPDDR (low-power DDR) allows it to be faster? And, by faster, do you mean lower latency? higher clock rates -> higher throughput? This is unintuitive to me.
My impression is that lower power means that you can't sustain higher clocks as readily (in fact, when overclocking RAM, it's common to increase voltage in the interest of stability).
I can't find anything about CAS latencies for LPDDR DIMMs.
edit: to clarify: when overclocking RAM, your two options are either increase voltage or increase timings, as if you want to sustain higher speeds, you need to either charge your capacitors faster, or wait more cycles for them to be charged.
https://cdn.arstechnica.net/wp-content/uploads/2020/09/tiger...
By faster I mean higher throughput at similar latency, achieved by higher clock rates. And it is indeed unintuitive as to how this can be done while using less power than standard DDR.
My understanding is that it's down to two major factors:
1. JEDEC has iterated on the LPDDR standards much more rapidly. DDR4 and LPDDR3 both hit the market in 2012. But then LPDDR3e, LPDDR4, LPDDR4x, and LPDDR5 were all introduced before DDR5 was.
2. LPDDR isn't available on DIMMs, it's soldered only.
So given that most laptops sold by companies like Dell and Lenovo use soldered ram anyway, and that Intel and AMD both support LPDDR, then why are PC laptops with faster RAM so rare? I have no idea, maybe it costs a bit more and the manufacturers don't think they can market it as a benefit?
For consumers, the primary application that benefits from higher RAM bandwidth is real-time graphics rendering, and non-Apple PCs optimize for this by using discrete GPUs with their own onboard high-bandwidth memory.
