PaddyNewman

Spent yesterday because it was cold and wet, for the comparison its against my 4080 Palit to have it closer as a 4090 is just well out of scope and not really comparable, I don't have a 30 series anymore which would have been better! All single player tests were done for 5 mins on the same save point, settings defaults to "high" - 1080p and 1440p, vsync off, no cap on frame rate where possible, or 144 used if no uncap was available. No XeSS/DLSS was on and RTX was disabled. Windows forced to allow games to handle 3D, rather than per application settings which I have currently to make it fair. CPU is a 7800X3D. FPS monitor - https://store.steampowered.com/app/966610/FPS_Monitor__hardware_ingame__desktop_overlays/ Games and results: Cyberpunk 2077 Nvidia 1080p: 90ish fps when shooting, about 90-100fps when driving at high speeds. 1440p: Only minor drops, 80s for shooting and 90+ when driving. Intel 1080p: Between 55 and 70 depending on action. Driving at fast speeds with motion blur on kept it in the high 60s. Full grenade explosions and gunfire dropped it down to the 40-45 region but realistically, 50-60fps is acceptable, it was higher on average but my save is mid gunfight! 1440p: This tanked it somewhat, average of around 30-40, highs of potentially 60 when driving with blur. ---------------------------------------------------------------- Destiny 2 Nvidia 1080p: 260fps in orbit, tower was 100-120fps depending on loading area, 140-160fps in Javelin4 no real obvious change outside to inside. 1440p: 180fps in orbit, tower about 100 give or take 5 either side, 130fps in Jav4. Intel 1080p: Orbit talking displayed 190fps, but in reality its showing nothing, Tower was about 70fps, crucible on Javelin4 seeing high 100 inside, 110 outside. 1440p: Orbit 160fps, Tower was around 60fps fluttering to 65 every now and then, crucible Jav4 was about 100 all around, no real dips. ---------------------------------------------------------------- Grand Theft Auto 5 Nvidia 1080p: Single player about 110fps, big explosions and like are dropping it to 90s, multiplayer was about 100+ on high, ultra took it to mmid 90s. 1440p: Same as above minus about 15fps for each, not noticeable but the numbers say different! Ultra was a bit sketchy into the low 70s. Intel 1080p: Single player about 85fps at best, dropping to medium and MSAA off it hit 90 when driving but any large action was ... oof. Multiplayer about the same however was purely as I drive a specific route every time and it was about 65-70fps on high. 75-80 on medium. Ultra is not. 1440p: As above but all down significantly, on high this runs like dirt, 45fps when driving and some severe dips when using rocket launchers on innocent civilians and cars. Seeing 40 online was a surprise. On medium it didn't change much, the resolution felt like it was taking too much. ---------------------------------------------------------------- Red Dead Redemption 2 Nvidia 1080p: Not far off GTA5, if anything it was a few fps higher but wouldn't notice, about 115-125 when riding. 1440p: About 100fps all round. Intel 1080p: About the same as GTA5 to be honest, but it was more stable around 60fps but nowhere near as much needed to render as its just tree tessellation and open skies. Highest was about 70fps but dramatically smoother than GTA5 1440p: Not as much of a dip, but again little to render compared, seeing a solid 50fps. ---------------------------------------------------------------- Call of Duty Nvidia 1080p: 160fps menu dipping to 120 when in game, lowest I saw was 90 but it was a random map, couldn't control that in multiplayer. 1440p: 150fps menu but around 90-100 in game. Intel 1080p: It works fine for the Black Ops 6 multiplayer, its a little rough on high and you do get shafted on fps with all the warfare going on with grenades and it almost starts artifacting when the screen is overwhelmed. Still a comfy 60-70fps on high. 1440p: 1440p seems to be this cards weakness, it really screwed this one up, saw 55-60fps routinely but the second you get a screen full, it really tanks and the only game to close with a graphical error! ---------------------------------------------------------------- Sea of Thieves Nvidia 1080p: Unsure if server problems but I had about 60-80fps when sailing. 1440p: Same as above, so perhaps erroneous game choice! Intel 1080p: Again, unsure whether its the servers at this point, but this was almost unplayable for the most part, its got fairly realistic seas and it was choppier than anything. Was seeing 35-40fps just sailing from Cutlass Cay direct north. 1440p: As above really, it was about 35-40. ---------------------------------------------------------------- Hogwarts Legacy Nvidia 1080p: 110 fps on the run to Hogsmeade, dueling is around 100fps. 1440p: 80 on a run, 80-85 when dueling. Intel 1080p: 70fps running Hogwarts to Hogsmeade, dueling class in castle is about 70fps too. 1440p: About 45fps running between Hogwarts and Hogsmeade, dueling was about 40, dipped quite a lot when fast casting, but nothing below 35, playable for most. ---------------------------------------------------------------- Helldivers 2 Nvidia 1080p: 160fps all round, close quarters is fine but obviously dips a little, only by 10 or so FPS. 1440p: Getting about 130fps planet side and dips to 100 when calling in close quarters nukes. Intel 1080p: Between 60 and 80 depending on enemy density, do get a little low when you call in close quarters, but really it never hit less than 60 and remained playable. 1440p: 55-60fps on planet, 70fps in dropship, close quarters call ins are just a nuke to this card. Seeing drops to 20 and almost felt like tearing when you try to turn mid explosion. Only on 1440p though. Thoughts - I would have one for that price if I was running a 16xx/10xx series Nvidia and even a 20xx series and it was aging badly. If a 4070Ti/4080 is well out of reach or you don't really need the beefiest card as you play decently optimised games... this is a STEAL to play games at sensible FPS. I did not want to test at 4K as it seems like an absolute waste of time, not many people wanting to get a cheap card, then get a 4K 120Hz screen... I'd love to have tested against a 6800/7700 card but the wife might get concerned for my health if I sit on the floor with PCs in parts even more than I do... Worth noting some of my results seem to vary widely with other results and tests. I can only imagine that me turning off all the "aids" doesn't help. They are what I saw and I trust FPS monitor

Mine has arrived this morning, I'll fit tonight, I imagine it wont stay in for long but will be good in the backup PC. I've only got a handful of games to test with and its against a 4090 so it'll be nowhere close, but be good to see how high it can push and maintain 60fps!

Make sure guest network only allows the absolute bare minimum port wise, TCP 80/443 filtered, DNS only to your chosen provider or your local DNS (ideally one that doesn't support ECH) and NTP so they can get time. Maybe if you are nice, push notification ports, but I am not that nice. If they get through that, its filtering.

Didn't notice they were that cheap, I've grabbed the TITAN OC from AWD, will swap out my card when it arrives. Always worth having a spare card for emergencies!

Had a port drop to 100full? Cable failure etc? If you plug in directly to your router with a known good device and cable and get the speeds, I'd be cable checking. 95 is what I'd expect with a 100full connection.

I imagine its an uphill and quite unwinnable battle with pinning. I think, and I'll have to test, a hardcoded proxy is fine, but transparent is just a no. I mean, Google is Google, you will conform etc, super power of the internet and all that jazz. It's incredibly annoying from an ISP side that really wants to push decryption though so I feel your pain!

They are the standard ports for Google push, unlikely have bearing on the 'internet' side but worth opening to the Google endpoints however.. counterproductive but exclude www.Google.Com From your decrypt, against this one device to test. Android is built on Google, www.google.com is pinned in that OS, best of luck and sorry for badly formatted post. It kept hyperlinking

May be late, but a customer just came to us with this, TCP 9441 was being requested to Amazon ranges, specifically 15.236.140.255 for me. Confirmed working.

But you can see other traffic on that device? Very odd... I'd be capturing the traffic on the firewall as it suggests its only 443, but you never know..

Can you see the live logs in Netsweeper?

In Netsweeper, add a decrypt bypass: decrypt://izettle.com - allow decrypt://izettlepay.com - allow Should sort it.

I'm sure it was mentioned in another thread but BYOD is not managed by you, hammer the filtering policy closed so it's basic web access (student but tighter, maybe office/Google email and generic education sites) and if it's too much hassle just ban BYOD. You'll get far less people connecting if you say 'you just install this certificate and extension/MSI so we can watch you' Only real problem you'll get is NHS staff, at which point they should use their 4G dongle/you pop em on the staff range temporarily.

Right now, my phone only on SL router. 44ms is my median over last 10 mins. Current 66ms. Now 71ms. 51 65 69 81 70 51 57 etc.. Its pretty low overnight as you'd expect with lower usage over this cell, but it ramps every day. It functions fine, but obviously you start pushing heavy traffic down it like I do every night and it is as you would expect, annoyingly one machine doing Windows updates can saturate the entire possible throughput too... and I do them on SL as my 8Mb ADSL is just dire for feature updates!

Either reserve/VLAN off those devices... or consider going sort of the other way. Authenticate your Chromebooks, get them audited instantly and move them from a default policy, can decrypt them properly then. Can't authenticate iPads with ease, but Chromebooks are 2 second job (as long as the Netsweeper service is exposed correctly)

I mean you could query LGfL for a potential price (I work there, so would say that) but I've used Starlink for over 2 years, nearly 3. It's not 'perfect' and I'm seeing a lot worse connection these days but it serves a purpose. I'd take their numbers with, well between a pinch of salt and a gritting lorry.

Biased but Netsweeper is the easiest filtering platform to use, I've found all the others remarkably complex or too basic. Wondering when you would need to filter on MAC address though. What's your aim, as you can definitely do what you need in Netsweeper, and happy to help if you want to reach out directly.

Yes, that sounds kinda loop-esque. If you have decent switching, it might be complaining about excessive broadcasts.

Again, non community, but if you have an LGfL feed, we can likely provide you with Senso!

Ah ok On your browser, if you go to Google and search for fish, top left in the address bar to left of http://www.google.com/whatever, should be able to see something that looks like o- with -o under it, or the word Https, right click (I think) that and click connection is secure, view certificateinfo. If it's GTS or Google trust services, it's not got decryption on, which means it's impossible to see past the .com part. Give it a whirl

Swapping IPs might help, however I feel your pain. My home is rural with a Starlink and my IP is known from the Berlin PoP for Starlink so I get German everything. Fine for the most part until you want the price of a phone, euros, or you watch twitch and get ads for German shoe makers with full German dub. SL suggest it's not a fault with the service. I agree, but I imagine it's just a pain to change it. I imagine they could, temporarily, NAT you to something else and see if it goes away.

Is there decryption on transparent? What certificate is applied to the search engine of your choice. Never used safety net, so absolutely blind, but is it a filter and monitor. Filtering isn't typically an alerting tool, more...reporting, from me at least.

Just by any chance are you an LGfL school? You might be entitled to free licences.

Oh I think I get you, physically taking the content of the whole page and running that through a lookup, rather than relying on URI. Similar to doing a quick wget.. Interesting, even with compression that's going to consume storage, but I guess if you log the bad and bin the stuff considered safe... I might actually give that a pop at home.

Forgive me if I'm wrong, but this is just a web filtering setup. You've just described a proxy based web filter to me. That's going to capture the URI and the string, but won't have the context, content or ability to read inputs, just what is fed to the proxy.

If you are keylogging, you need to manage the machines, and to effectively do that, you NEED a client on a managed school device. If you allow unmanaged devices on your internet, your obligation is to make sure that feed is filtered and firewalled. Either allow them and accept the short fall there which is "they are not school owned, therefore we cannot do more" or terminate BYOD personal devices. Try their 4G, get to all of the internet. TLDR you cant do anything on your internet feed to keylog, its client based. You can SSL inspect and see the full URI/search/but typed content is not usually captured without a client.