Flea Yatsenko
-
Posts
512 -
Joined
-
Last visited
Content Type
Forums
Blogs
Knowledge Base
Posts posted by Flea Yatsenko
-
-
Crossfire and SLI need direct access, which means exclusive full screen. Borderless window full screen doesn't cut it. Something to do with bypassing the OS compositor or something. Actually they sort of got it working in SLI with windowed screens, but it depends on the drivers and needs a lot more work, and sometimes it doesn't work as well as full screen SLI. But AMD and NVIDIA both sort of gave up on Crossfire and SLI. There was a time they both though multiple mid-range cards was the way to go forward, but they both changed their minds. Now we have 700mm^2+ monster GPU dies instead of two or three 300mm^2 ones.
-
Crossfire and SLI (multiple cards working together) only used to work on consumer drivers with full screen windows. I don't know if it's still the case. I remember having a Crossfire setup long ago when they removed full screen support and it really messed me up. But Crossfire and SLI are pretty much dead now so it's not a big deal.
-
It would be nice to be able to send a message with the redelivery that explains what's new. If someone gets a new update they won't even know what's new. If they don't think there's a problem with the old one, they won't go through the hassle of replacing the old version with the new. But I feel like a feature like that would get abused even more than this one has. Which is too bad because it would be really great for customers and merchants.
-
This needs to be fixed, it's driving away a lot of customers. How many people do you think spend the time to come to the forum and ask for help, or who know how to clear the cookies and that that's the problem? And how many people do you think just close the browser and go play something else? Considering LL is shifting revenue from land payments to commission, LL and merchants are missing out on a lot of income from this.
If LL wants to shift from their revenue coming from SLMP instead of land tier they need to ensure SLMP is running at 100% as much as possible, and driving away as little customers as possible too. Land tier always brings in revenue for LL no matter how broken or empty the grid is. If the grid is in bad shape and SLMP is broken for a lot of people, LL won't make as much revenue. Both situations hurt merchants, but the shift to SLMP revenue for LL can really harm their revenue. Just saying.
I sure would hate to see the plans of land prices coming down because they are being replaced with SLMP commission fail because SLMP doesn't work for a lot of people and the SLMP revenue doesn't allow LL to lower the price of land. Please don't let the situation become you've increased the commission on SLMP sales to lower land costs but SLMP is broken most of the time so SLMP sales never increased enough to lower land prices, so we end up with higher commission on SLMP sales and land prices don't come down. I realize there's several factors playing into this, but this August is shaping up to be one of the slowest Augusts I've had in like 4 or 5 years, and I'm not the only merchant I know who has been having below normal sales. I don't mean to be rude, I really appreciate everything LL does.
-
Thank you so much for this.
-
It takes two whole seconds for ANYTHING to show up on the page, that's before the HTML is loaded so the browser knows which images to fetch. After the first two seconds for FCP then the browser starts to download the images. Largest Contentful Paint is the time when the largest DOM element on the page is displayed. The first two seconds of that load for the page to start has nothing to do with the images, it's the time it takes a (fresh) browser to render anything on the page, including any HTML. Usually because it's waiting on the network, possibly because of excessive redirects.
FCP is not when the entire page and every image is loaded, it's when the browser displays something, anything, even if it's just the first bit of HTML that hasn't even been styled with CSS yet.
-
4 hours ago, bigmoe Whitfield said:
why? no point in fixing it currently, plus I'm gonna go out on a limb and maybe just maybe think that this is build around their backend and is not as straight forward and rip and rebuild and that's just gonna be that.
SLMP is basically an eCommerce site, but LL is the mediator, it's a lot like eBay. A ton of research and improvements with increasing sales has been done since the last time SLMP was ever touched. And go compare it to another eCommerce site like Amazon or eBay. Amazon and eBay both load almost instantly on my computer, and when you log in you see what you've recently purchased and recommendations on what you might like. It's a clean design, no extra stuff, no fancy colored banners, just a focus on the products for sale.
Now go to SLMP, the page takes a long time to load (anything over 300ms is basically death for an eCommerce website), a quarter of the width of the page is just a bunch of categories which could be used to display relevant products instead. The top part is a carousel (which no one uses anymore) suggesting what I assume are promoted goods. Under that is another carousel with promoted goods merchants pay to have promoted. Below that is a few listings of what people are "buying now". Which, by the way, doesn't even use remotely standard CSS like a flex box to display and instead uses floats.
SLMP needs a huge redesign. These out-dated carousels need to go, it needs to load faster, it needs to show similar items to what you purchased, make recommendations, and do a better job shoving products in your face that you might like. Look at all the categories a place like Amazon or eBay has when you load the home page logged in. Recommended categories, recently viewed items I looked at but didn't buy (so I can be reminded and maybe they can get a sale out of me), a special deal of the day, recommended videos, items related to stuff I've looked at, more items I might like because of my buying and viewing history, a few little seasonal things that aren't really tailored to me after all the stuff that's been tailored to me. A ton of recommendations for things based on categories of products I view or buy. EBay is similar just not as good.
But my point is that an SLMP redesign isn't just about the way the site looks. It's about how it gets products to people and the site helps them find relevant stuff they want to buy. The home page doesn't do this at all, it's why everyone depends on search and why no one can find what they need most of the time, and why there's so many problems with search. Search is what people use when the home page isn't giving them what they need. Going to eBay or Amazon I can find stuff I'd like to buy just by opening up the home page. Marketplace is showing me some beach stuff and a ton of female clothes I'll never buy. Getting search to work on a site of user generated content that earns the creator money is pretty much impossible to get working very well. It's always going to be an uphill battle. But the home page should show you stuff you'd like to buy that you'd like that interests you. Right now it just shows you what other people are buying and some stuff merchants paid to promote to anyone (completely un-targeted advertising). SLMP needs someone with eCommerce experience to revamp the whole thing and give visitors to the home page something they want to buy before they even search.
Marketplace's performance is absolutely abysmal. It doesn't even work on the mobile version. I'm not trying to be mean to LL or the SLMP team by any means, what I'm asking for is a lot and it would require something from the outside. But I can almost guarantee you they'd make a lot more money increasing sales by showing relevant products to people on a fast loading website than they would getting paid for featured listings on the home page.
-
2
-
1
-
-
On 7/16/2021 at 9:14 AM, Qie Niangao said:
I wonder, though, if there might be another solution, preserving the 256x256 unit, while permitting those individually addressable units to be combined for the purposes of processing. So while X and Y dimensions are still in the same range, the stuff on those units are handled by the same threads and share the same memory. Obviously it's not at all a trivially simple solution, but it may be simpler, and achieve folks' "larger region" objectives (e.g., longer-range travel with fewer handoffs), avoid some of the compatibility problems for existing user-generated content, and sidestep the inherent floating point precision problems of larger in-region coordinates.
You are basically just advocating for updates to sim crossing performance. I don't think we're ever going to see the current Second Life have regions larger than 256x256. The fact they created homesteads which still are 256x256 should give you insight to how difficult working with this is.
The only chance they have for different sim sizes is to make smaller sim sizes, but it would just be pretending to divide a 256x256 simulator up into smaller sims. Which wouldn't really be worth the time or effort to do because you'd still be constricted by other people on the sim. i.e. there are 4 people with 128x128 fake mini sims and you want to expand your 128x128 to 128x256, well too bad someone else has that quarter. So they'd have to be moved or something. I really want to see it happen and I thought the move to AWS would make things easier since they are virtual private servers so you can add CPU cores, RAM, etc at will, but the code won't allow it. The only way we're going to see sims "expand" is raised limits for maximum number of avs on a sim, higher land impact capacity, etc, probably at a higher cost.
The sim size is not changing, maybe we could some day see sims that have different priorities or levels you can buy (i.e. script performance, number of avatars, prim capacity, etc). But I think it's always going to be 256x256 until someone at LL really wants a big project. And to be honest, if 256x256 is baked so far into SL, going around and changing it is going to make a ton of bugs in other things. I actually wonder if the issues making these huge changes to support sims other than 256x256 would be worth it, because if it's that deep into it changing it is going to break a lot of things no one at LL can foresee.
-
As a Linux user myself, having a third party take care of Linux support makes the most sense. LL has more important things to do yet a Linux port is helpful to a good amount of members of the community. I think if you are motivated enough to install Linux that you won't have problems figuring out lots of SL on your own.
I really like the floater guidebook, it's a really nice concept that'll help. There should be some focus on building though. Looking at that a new user wouldn't know you can build. I would think even just mentioning something about the user generated content and building would be a good start. Maybe someone would TP around to a bunch of empty sims and think there's nothing there, but if people were diverted to a building tutorial and sandbox if they wanted it would help retain users who want to make things.
SL is unique because of the user generated content. It's a double edged sword. In a typical MMO, the developers add more content and servers when they need to. SL has people buying servers so they can just make stuff. MMOs are always filled up because they build to meet demand. SL is empty a lot because people are building trying to attract people. I think a huge problem new users have is going to explore and finding empty places. They need more reasons to stay in SL than just being social and looking good.
-
1
-
-
On 6/28/2021 at 1:26 PM, Sapphire Dakota said:
Basically, the price ratio of land space to prims is weird to me. What's the server cost based on, size or prims allowed? Either way, the math is odd.
An example? I'm getting 1406 prims on my current rented parcel for the Linden equivalent of $4/week aka $17.33/month. By that math, 5K prims should cost me $61.63 per month, even under what I'd pay at $99.
There is a deep rooted thing in second life servers that restricts it to 256x256m. You may notice it's a power of two in each dimension, for a reason. One that probably doesn't need to exist anymore but a lot of stuff would make the assumption that the dimensions are under 256m. I understand the homestead thing, some sim builds are a lot more dense than others. For a store that sells large objects, a homestead with a rezzer is a great set up. I would love to see different sized sims, if only for the fact that you could have a smaller parcel to yourself on a whole sim. I just need half of a homestead, but I still have to share the other half with another renter. Which comes with all that exciting stuff, like him changing experience settings for the whole sim on accident and all that.
As for AWS being more expensive per month, it's not accounting for the fact that LL's hardware was getting old in their data center. By the time you start upgrading the hardware, price gets out of control very quickly. Colo is also kind of restricting for growth, what happens when you outgrow your datacenter? Now LL is responsible for managing their servers and providing enough to meet demand. With AWS they can just order more servers and Amazon does it all. I am not a fan of AWS but SL does not scale very well, with sims needing their own dedicated resources no matter how much they're being used. IIRC there was a problem with there not being enough land available to buy because they didn't have the servers. From a business perspective, if you can't meet demand you are just losing money. I mean, think of it this way, it's $229 for a private region, you run out of capacity and turn away 100 people because you don't have the resources to provide them with sims. Not only is that the $2,290 you lose in the first month, but they are recurring customers. You are losing tens of thousands of dollars a year because you can't meet demand. It makes sense to spend more on servers if you can meet demand.
The older hardware they were running on their servers also becomes a liability. If you have components dying in your servers, not only do you have mad customers who have their sims they are spending money on unavailable, but you have to spend the money to upgrade them. You also have to pay someone to take care of that stuff.
While AWS is more expensive, I think they made the move from a growth perspective.
-
2
-
-
Just slapping multi-thread on an old engine isn't going to work. You need to make major changes instead of just making some parts of it multi-threaded. Otherwise you just keep profiling and chasing the next single thread bottleneck. Multi-thread has a lot of potential to make SL faster, but it's not like you can just make parts of existing SL multi-threaded and have it be some sort of cure all magic performance boost.
Think of it this way. the SL client is a huge amount of single threaded code. So you look at the slowest parts you can find and multi-thread them. But everything else is still single threaded. So maybe you'll speed up a few specific things but everything you didn't touch will stay the same. Meanwhile, you've just re-written a bunch of code that worked fine and risked adding a ton of bugs with new code that's much more complex. Which leads to situations like, "performance increased up to 15%" and people mad things that worked fine were broken.
But SL's greatest strength is also it's biggest weakness. There is near endless content on SL for anyone, more than any other virtual world by far. But some of that content has been around for a very long time, and deprecating it can cause problems. Realistically though, people are upset over losing something they probably spent a dollar or two in real world money on. If there is demand for a deprecated product someone will fill in the supply void to meet the new demand.
If LL was to do something with the client, they'd need to do a full engine rebuild from the ground up. I've always been a fan of splitting clients into a simplified, high performance "regular user" client that doesn't allow building, but still lets you customize your AV, and a "content creator" client which is basically exactly what SL is now. But when I've mentioned it before people got quite bothered by it. I don't know why, it's a chance to rebuild a new client that's much easier on new users that runs a lot better. 100% everything new users need to stay in SL. And when you want to build you just fire up the content creator client and have all the tools and features you always had. I'd wager the vast majority of people in SL only need an interface to teleport around and change their avatar.
-
2
-
1
-
-
I've been noticing problems with other websites that use AWS (just like SL uses) lately, including sites just flat out not loading. A few days ago honda.com wouldn't even load for me, the connection was getting lost in AWS. It's fixed now though. I don't know what's been going on but the entire internet has been acting oddly for me for the last few weeks. Not trying to defend SL, usually this time of year is lower traffic because of the weather and things run smoother than normal. My sales (and a friend's) have been well below average for this time of year as well, I think it has a lot to do with people having slow loading textures on demo items when they rez and them just losing interest and moving on. I would have thought things would be going well with a large portion of SL's users being in the USA and most of USA opening back up after lockdowns, with people rushing to go outside. Basically less users on the internet and SL should mean less problems like this. But it's clearly not happening.
I've been having problems with other websites as well as SL lately. I have no idea what's going on but I think it's larger than SL.
-
On 6/20/2021 at 11:40 PM, Coffee Pancake said:
That's still a thing .. Built a huge rocket ship for Catznip this year. Although I will say the foot traffic is noticeably lower than previous years.
A lot of US COVID restrictions are being lifted and a lot of people are going outside for the first time in a while. Last Summer we were all pretty much locked down, but a lot of states are either almost completely unlocked or are very close. It's not just SL traffic that's been down, my RL internet stuff is way down too, but it's slowly recovering. I think all these cyber-attacks have been affecting things negatively too.
-
15 hours ago, Rowan Amore said:
The example I sent with my jira. No where on that page is Sage Mellow.
I think understand why this is happening, it's because "sage" and "mellow" are both common terms used all over. It makes sense. Normally when you have this problem you can put your search in quotes, so instead of Sage Mellow being searched you do "Sage Mellow" (with the quotes). But it had no effect. But looking at the results on my end, it looks like it's just searching for "sage" and "mellow" and merging the results for the two instead of finding everything that matches sage AND mellow. I don't know if that makes sense, but it's not looking for results and saying "This contains 'sage' and 'mellow', rank it higher", it's just taking everything that matches "sage" and everything that matches "mellow" it's like it's searching for Sage OR Mellow instead of Sage AND Mellow.
I'm assuming the same thing is happening with "Ravenglass Rentals", instead of searching for things with ravenglass and rentals it's searching for ravenglass and searching for rentals then merging the results, so things that just match "rentals" are thrown into the results.
I don't know what software LL is using for this, I use ElasticSearch and have experience with it and this kind of thing is usually a really simple mistake. I've been lurking here because search functioning properly has a huge effect on everyone, especially merchants who are trying to get their relevant products to people who are looking for them.
-
On 6/5/2021 at 10:08 PM, Paul Hexem said:
Is single sign on related to the redirect bug that keeps happening, too?
That bug has been around far longer than single sign on. I keep my dashboard as a pinned tab and every once and a while it just heads into a redirect error.
-
There have been intermittent problems for the last week or so, but it's really bad right now. Sales have been well below average year over year and far below expectations. I'm sure they know everything is down. This costs LL and merchants a lot of money when everything is broken like this, especially with LL shifting revenue from land tier to marketplace and L$ transactions. If the websites are down there are no transactions and no one makes any money. I keep checking to see if there's any info from LL about what's going on but they're awfully quiet. My RL work is taking a huge nose dive because of the way the economy is going and I really need LL to pull through here and provide a reliable ecommerce platform that everyone can depend on.
-
Why isn't this places thing a bigger deal? It's very helpful and a great way to tell people about your land/store/etc. I really like it, honestly it's one of the nicer things I've seen with LL website based stuff.
On topic, there are some good use cases for this. My in world store and SLMP demo rezzing depends on opening up a web page. It would be amazing to be able to automatically open up that page, or something other than a dialog box requesting to open the page in a browser. But I understand a lot of concerns in this thread. In fact I think the dialog isn't that good for my situation already because it has been abused. I open up a dialog to request to open a web page because it's essential functionality for my store. But most people use dialogs to spam people when they enter their land to join groups and all sorts of stuff. It's pretty obvious to me, in my position, that opening up a web page would be the same idea. And of course if it's any web page malicious people will be storing IPs, it would be trivial to have a script open up a URL with the AV's name and key as URL parameters then have the server get the IP address. LL would have to run their browser through a proxy server kind of like cloudflare does to hide the IP of the user.
-
On 5/2/2021 at 7:15 PM, Tazzie Tuque said:
Good grief! MP is hard enough to get things listed on now as it is, what with all the frequent maintenance and errors when trying to list.. I cannot imagine the fiasco if they re:booted it and everyone had to start all over again.. as a merchant, there is no way in hell I would have the time to do something like that.. and I am pretty sure I am not alone.
I don't want to see marketplace rebooted and have to see everything relisted. I have a lot of products and there's people who have way, way more than me. I also see the value in old products, I've tried to clean things up personally and I've had people message me requesting things I thought were too old to be useful but they really wanted them.
But the reason why SLMP has so many technical problems is because it's so big. When you make a listing on market place, that number at the end of the URL is the ID. Assuming it works like most database ids, it started at a low number and increments by one every time a new row (listing) is added. Looking at a listing I made a few days ago that number is around 22 million. Obviously it's only an estimate and maybe LL did something like start the number high. But my point is there's very likely 10s of millions of things on SLMP LL has to maintain to keep the SLMP website functioning. It's a pretty huge burden honestly. Granted the size of SLMP and the issues the size causes I think LL is very generous on what it takes to start up a store and start creating listings.
-
This is a lot easier for LL to do now that they are in the cloud. I've seen it discussed here several times that sizes besides 256x256 would be possible and welcomed. You also have a benefit of having smaller ones too. It would make it easier to expand, instead of managing a 256x256 sim you could get a 128x128 and when you need more prims just pay for more and have your sim expanded instead of paying for a full size sim and trying to sublet it. But I think this is one of their lower priority issues. Seeing LL's latest blog posts they are concerned with improving performance and getting new users to stay in SL. More people in SL means more income for LL (and merchants) which means more growth. But it's a lot easier for LL to do this now they are in the cloud and not tied to fixed hardware with fixed specs. I think it will come but it's not going to be a huge priority for them. They have stuff to fix and I'd rather see them fix it as much as I would love my own 128x256 private sim I could expand as my business grows.
-
1
-
-
4 minutes ago, Mollymews said:
is a interesting question this
i read in a User Group report recently (over on Inara Pey blog as I remember) Linden may also go with more than one render API in the future, if/when they do a major rewrite of the viewer. Like Metal for Apple, DirectX for Microsoft. Maybe?? also continue with OpenGL for Linux and Potato. So three viewer versions for each release maybe
Linden do have do something for Apple users going forward, so some kind of quite major viewer rendering rewrite has to happen at some point
Metal and DX12 are all based on Vulkan. Sadly I think we are heading into a future of fragmentation, Vulkan should be supported on all major platforms, even if it requires some third party libraries to go from Vulkan to Metal or Vulkan to DX12. It just depends if Intel, Nvidia, and AMD are going to support Vulkan properly going into the future or not. I hope they do, having different APIs for OSX, Windows, and Linux is going to make a huge mess of things and distract from features like having a proper Android and iOS client.
OpenGL was never a first choice for graphics, it's focused on accuracy, not speed. But it was chosen by many projects because it's easy to get it running on multiple platforms.
These new APIs will allow SL to use more than one core effectively, but it's going to take a lot of work, they'll be better off starting from nothing then trying to get it working in the current engine. They did a really good job with Sansar though, I know they can handle it fine.
-
1
-
-
I thought the major reason for moving to the cloud was that LL's bandwidth from their servers wasn't that great. It also means LL doesn't have to maintain the server hardware, and AWS is a lot more resilient if something goes wrong (like a hard drive dies). Uplift was never going to be a cure all. But I think it helped with a few things.
-
1
-
1
-
-
Multithreading and Vulkan has been brought up a lot before. IIRC there is a problem where network fetching or something causes a single thread bottleneck. It was something unique to SL having to fetch things off the network. It's in this thread, specifically this post:
LL would basically have to re-write the engine, but they'd still be held back by the existing content, like using normal and specularity instead of PBR, not including massively unoptimized content that people have created. And the content on SL is the most valuable part of the platform, we've seen lots of technically superior versions of SL come and go, because they don't have the content and stuff to buy that SL has. I feel like a lot of people take for granted you can go to the marketplace and buy pretty much anything you want. It's there. It's amazing. LL built a fantastic ecosystem. But it takes a long time to build such a good ecosystem and the server and viewer software show its age. It would be a very huge undertaking to make SL multi-threaded. But x86 CPUs are done getting much faster per core and are just adding more cores.
A large part of optimization with SL comes from not being able to prepare the world. In a regular game, everything is set in stone and it's optimized assuming nothing will change. So you can take short cuts and make it run better. SL is all off the network, and it's always changing.
-
1
-
-
1 hour ago, Coffee Pancake said:
That was mostly just a UI rebuild by people who didn't use SL. The concept of a side bar was great, the implementation forced everything to use it & bottlenecked everything everyone did.
This time round, lets start with a new asset fetch-decode-vram-render pipeline, something that might make use of all these cores we have laying about collecting dust, bit of vulkan with a little RTX on the side so our fancy pants GPUs have something to do.
Modern PC's are wasted on SL.
I don't know what made me search for "Vulkan" in the forums, but I found this active thread, and I think you're right on the money with this. Vulkan can multi-thread rendering a lot better than regular OpenGL. I kind of went down a rabbit hole, but I found this regarding performance http://wiki.secondlife.com/wiki/Culling. The edit history is quite old, I don't know how much is actually relevant.
Some of it may be outdated, but the Performance section has draw calls being a bottleneck, and the wiki considers adding more triangles to draw calls. But OpenGL's draw call performance is much lower than Vulkan, I've seen benchmarks with Vulkan being 2 to almost 4 times faster when benchmarking draw calls, with reduced CPU usage.
I wonder what the draw call bottleneck looks like for SL. It's dynamically loading content from the web, so mesh is slowly trickling in, but I don't know if it would make a difference. I don't know enough about this stuff. I'd think it has a lot of potential to increase performance. It'd be good to get the SL viewers more multi-thread friendly, especially when phones usually have 8 very weak CPU cores to use, as a long term goal.
-
I filed a ticket, but just in case some Lindens are around and investigating, I dumped the network requests when trying to log in and it's just stuck redirecting to https://secondlife.com/auth/oid_end.php?redirect=https%3A%2F%2Fsecondlife.com%2Findex.php&janrain_nonce=2020-12-21T21%3A37%3A35ZJr4mMF&openid.mode=cancel&openid.ns=http%3A%2F%2Fspecs.openid.net%2Fauth%2F2.0
over and over again, until the browser gives up.
Facebook
Instagram
Twitter
YouTube
Flickr
Is Second Life too big now?
in General Discussion Forum
Posted
This is it, perfectly. Even the SL's hay day, most sims were empty. You can look at how many sims there were and how much fewer there are now. But how many of those were empty sims people gave up on? How many are still empty?
How many people search for sims they have interest in (like a cyberpunk RP), only to find an empty sim? The sim count doesn't really reflect how well SL is doing. I would even argue it'd be better off with far fewer sims that had more people in them at all times.
I think a large part of trying to shift LL revenue away from tier and towards marketplace was because they knew the model of charging this much money for a sim that's usually empty isn't worth it to most people.
SL's problem is their user to server ratio is bad, You are more likely to randomly teleport into a completely empty sim than you are one with someone on it. But the best solution is to get more people to fill up those sims, but sometimes you can't do that. Just imagine if other games were like that, like Overwatch or TF2, where you keep having to try entering servers thta