Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
Rafuino
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
21 ms
·
151.
▲
MySQL and Memory: A Love Story (Part 2)
(lefred.be)
2 points
by
Rafuino
7y ago
|
0 comments
152.
▲
by
Rafuino
7y ago
Startups would need to have sophisticated accounting and tax legal teams to do what large corporations do. Ideally we close these loopholes so it's not a strategic advantage for larger corporations. https://www.investopedia.
153.
▲
An Interview with AMD's Forrest Norrod: Naples, Rome, Milan, & Genoa
(anandtech.com)
1 points
by
Rafuino
7y ago
|
0 comments
154.
▲
by
Rafuino
7y ago
The article OP linked to has 24 trackers blocked by uBlock Origin, where the original source has 8 trackers blocked. Either way, get an ad blocker, I suppose, but we should read the original source wherever possible to support the actual j
155.
▲
by
Rafuino
7y ago
This is the actual report from Defense News. Why not link directly to it? https://www.defensenews.com/air/2019/06/12/the-pentagon-is-b...
156.
▲
On Databases and Non-Volatile Memory Technologies
(odbms.org)
3 points
by
Rafuino
7y ago
|
0 comments
157.
▲
by
Rafuino
7y ago
Not Oklahoma City, an Oklahoma city (lower case 'c'!). Actually, it's more of a tiny town... Tuttle has ~6000 people and OKC has >100x that.
158.
▲
by
Rafuino
7y ago
I searched the the paper and it doesn't seem to falsify what I linked to, but I'll have to dig deeper into the research. "Recent Intel systems" isn't specific enough.
159.
▲
by
Rafuino
7y ago
Better Intel page on the MDS vulnerability is here: https://www.intel.com/content/www/us/en/architecture-and-tec... . Interesting point: "MDS is addressed in hardware starting with select 8th and 9th
160.
▲
by
Rafuino
7y ago
Someone's gotta benchmark these and let us know
161.
▲
Memcached – Part II: The Volatile Benefit of Persistent Memory
(memcached.org)
6 points
by
Rafuino
7y ago
|
0 comments
162.
▲
by
Rafuino
7y ago
8 TLC drives could easily hit 3.3M IOPS before... i3 have 8 1.9TB NVMe SSDs max, right? A 2TB P4600(TLC NAND) has 610,000 random read IOPS max, so that's 4.8M IOPS right there. A 7.68TB P4320 has 427,000 random read IOPS max, so that
163.
▲
Oops, We Forgot to Build a Managed Kubernetes Service
(packet.com)
2 points
by
Rafuino
7y ago
|
0 comments
164.
▲
by
Rafuino
7y ago
My take on random read speed? Check out this paper from Lenovo testing the performance of Optane DC Persistent Memory on their servers: https://lenovopress.com/lp1083.pdf . There's another good paper from a team at UCSD
165.
▲
by
Rafuino
7y ago
Not OP but, yes, server vendors are shipping with Optane DC Persistent Memory, like Lenovo [1] and Supermicro [2], among others. Google also has a testing program [3, 4] with access to instances with 7TB of total memory. [1] https:/&
166.
▲
Memcached – The Volatile Benefit of Persistent Memory
(memcached.org)
1 points
by
Rafuino
7y ago
|
0 comments
167.
▲
by
Rafuino
7y ago
What do you use the 280GB 900P for in front of your 1TB drives? You're right the bottleneck is often somewhere else in the system once you have modern SSDs. The 280GB 900P is ~$0.89/GB [1] and 1TB EX920 is ~$0.15/GB [2] the
168.
▲
by
Rafuino
7y ago
They were the first to cut back supply. I'm guessing you work for them? https://www.techspot.com/news/79303-micron-follows-samsung-r... https://www.bloomberg.com/news/articles/2019-01-30&
169.
▲
by
Rafuino
7y ago
Has this pointer swizzling concept been adapted by any big data-focused databases? Would be interesting to see how this works with bigger pools of volatile and persistent memory (disclaimer: I work at Intel on Optane DC SSD/PM-related
170.
▲
by
Rafuino
7y ago
"As those natives who have more resources leave, those left behind may struggle to support churches, police athletic leagues, parent-teacher associations, and local businesses." Huh... churches are the first thing that come to min
171.
▲
by
Rafuino
7y ago
I think it's partially cultural and partially economic. I left my home state (which scored poorly in this Senate report) because I couldn't get a great education in the state system, nor were there any prospects for the kinds of
172.
▲
by
Rafuino
8y ago
What kind of storage do you use when chasing the 5-10ms improvements? I ask as someone working on super fast storage and memory and it's not always easy to find people who are aware of the differences in latency. I imagine you'r
173.
▲
by
Rafuino
8y ago
I'm not experienced with S-1 filings, but don't they usually include the share price targeted and amount they're looking to raise? It's blank up top on the document right now. When is that typically filled in?
174.
▲
by
Rafuino
8y ago
What I wonder is why Google spends tons of CPU resources to calculate routes that would take me much longer to get where I'm going.
175.
▲
by
Rafuino
8y ago
So, when is the FTC going to actually bring down the hammer on FB for violating the consent agreement? There's no way this was "unintentional." At $40,000 per user per day [1], even at just one day of violation, that's
176.
▲
by
Rafuino
8y ago
I can't wait for the CA law to come into effect. "The law, set to take effect next year, gives California residents the power to view the types of data companies collect from them, request that the data be deleted, and allows res
177.
▲
by
Rafuino
8y ago
Yep, I'm definitely familiar with the history coming from Intel :)
178.
▲
by
Rafuino
8y ago
Very cool. Had a busy day at work so couldn't check all the details!
179.
▲
by
Rafuino
8y ago
Yeah good point. Max and the various 9s that matter to your organization.
180.
▲
by
Rafuino
8y ago
Wow, that Netflix documentary is reallllly helping grow interest in F1! Enjoyed the long scroll down to see Haas as the winner. I assume the article controls for the physical location of the request (i.e. is Haas fastest because the autho
More ›