Gainward GeForce GTX 1630 Ghost

[ModEl4]

I don't understand why he reviewed this. there is no review of 1030, same should go for 1630

you can play chess on it i guess, but anything else is causing pain, and some people may enjoy this, but I wouldn't touch anything less than 1660S.

GTX 3030 based on the never released GA107 that was supposed to be a 3050 would have been a far better placeholder for the lower end until 4030.

There are a lot of people that are interested for a below $160 VGA product, me included.
Either they live in low/middle-income countries or are poor or they just don't want to pay much if they game casually or they just want to know in order to recommend it to their parents/friends or whatever.
The only problem is pricing with this product, no-one is surprised with the performance level being a so cut down iteration of TU117, if this product was at $119 it would be an alternative solution in the current market condition and at $99 the demand would be great.
Depending the games tested, it should be nearly 1.35X faster than a 5700G on 1080p Ultra and at least 1.1X faster in 1080p minimum settings.
Edit: actually it's more than the above, on average at least 1.4X faster on FHD Ultra and at least 1.32X on FHD min with same level CPU but tested without the older esport titles.

 

[Chrispy_]

I want to play games. My GT 1030 has been enough to play all of the games I've wanted to far, but it won't be enough for future games. It has been enough because I use low settings and a low resolution. My screen's resolution is 1360 x 768. A speed boost to around 2 to 2.5 times the power of the GT 1030 would be great. The RX 6400 would be even faster, but I need to know whether it would actually be slower than the 1630 because of the PCI-E 2.0 x4 limitation.

To be clear, I have a 2.0 x16 slot, but the RX 6400 has only four lanes., while the 1630 could use all 16 lanes.

AMD Radeon RX 6500 XT PCI-Express Scaling

The AMD Radeon RX 6500 XT comes with only a narrow PCI-Express x4 interface. In this article, we took a closer look at how performance is affected when running at PCI-Express 3.0; also included is a full set of data for the academically interesting setting of PCI-Express 2.0.

www.techpowerup.com

The RX 6400 is likely to also lose about one-third of its performance in a PCIe 2.0 slot, making it a close match for the GTX 1630 in your situation. Those charts show that the lower the resolution is, the bigger the performance drop from using a PCIe 2.0 slot!

Here's a new question for you - do you have cooling/airflow limitations, power supply limitations, or physical size limitations? What's your reasoning for wanting such a low-power card?

• If it's cooling - a 1060 founders edition or RX480 reference card are good candidates that dump all their heat outside the case and therefore shouldn't get hot+noisy.
• If it's power, there are plenty of QUIET GTX 1650 cards that do not require additional power plugs.
• If it's physical size and you must have a low-profile card, the Gigabyte 1630 low profile may actually be your best bet with a PCIe 2.0 slot.

There are plenty of people here who would be willing to suggest something - maybe worth starting a new forum thread for it with more details...

 

[looniam]

You linked PCI-Express SIG spec, which is nice if you're a motherboard maker, but my beef is with case compatibility and enclosures which is clearly the realm of the ATX spec. Cases are built to ATX/mATX/MITX standards, not PCI-SIG standards.

what? no. add in cards fall under pci-e sig. ATX specs the PSU only and most certainly not wikipedia for a source.

All that matters (and my original point) is that cards aren't supposed to peek out above the expansion bracket. That's the spec, whether you use ATX, PCIe, VESA, or anything else; All standards agree on that.

as above, there is one standard for add in cards and thats pcie-sig period. maybe look at them:

your info is clearly all wrong accounting for the space differences for no edge connector:

114.55-16.15 is 98.4 as illustrated here:

cases and their openings are (where you have your beef at):

i'll also point out how every schematic that has a bracket attached to the pcb shows the pcb is higher than the bracket. PCIE-SIG covers all that you are more than welcome to look at the latest ATX standards (attached) and see add in cards are no where to be found. read pages 14-15 and you'll see all the other related PCIE-SIG documents/standards.

problem is PCIE-SIG keeps all that behind a paywall and leads to what you are doing, relying on unreliable "popular wisdom" information. read the specs - there is a whole section that covers full/half height I/O brackets.

 

[defaultluser]

I want to play games. My GT 1030 has been enough to play all of the games I've wanted to far, but it won't be enough for future games. It has been enough because I use low settings and a low resolution. My screen's resolution is 1360 x 768. A speed boost to around 2 to 2.5 times the power of the GT 1030 would be great. The RX 6400 would be even faster, but I need to know whether it would actually be slower than the 1630 because of the PCI-E 2.0 x4 limitation.

To be clear, I have a 2.0 x16 slot, but the RX 6400 has only four lanes., while the 1630 could use all 16 lanes.

Yeah, then you will probably see similar performance to the 6400. Or, you can step thins up, and likely go significantly faster than teh 1630 with a full-fat 1650!

 

[W1zzard]

Added a test run at 1080p lowest possible settings in all games

Gainward GeForce GTX 1630 Ghost Review - Challenging the AMD RX 6400

The NVIDIA GeForce GTX 1630 launches today. The new card is targeted at the sub-$200 segment and goes head-to-head with AMD's Radeon RX 6400 and Intel's ARC A380. Unlike the Radeon, the GTX 1630 has support for the full PCIe x16 interface.

www.techpowerup.com

 

[Chrispy_]

what? no. add in cards fall under pci-e sig. ATX specs the PSU only and most certainly not wikipedia for a source.

as above, there is one standard for add in cards and thats pcie-sig period. maybe look at them:

problem is PCIE-SIG keeps all that behind a paywall and leads to what you are doing, relying on unreliable "popular wisdom" information. read the specs - there is a whole section that covers full/half height I/O brackets.

So what you're saying is that this 1630 might actually be compliant with the spec if you look at some paywalled info? Fine.

That's great, but it doesn't change a damn thing in the real-world because of that top-edge connector effectively negating the point of the spec even having a maximum height in the first place. Intel and the PCI-SIG can pedantically disagree by a few milimeters as much as they want, but the end result is still the same; The GPU won't fit in someone's case because of where the connector is oriented and you're saying that is omitted from the spec altogether?

What a dumpster fire of a spec if it imposes height restrictions but then fails to provide any guidelines for plugs and connectors that are mandatory for the card to operate, could be literally anywhere on a card (*cough* 3000-series FE cards) and completely violate those very locked-down height restrictions that the spec covered in so much detail.

If this were a mainstream gaming GPU I'd give it a pass; Gaming PC cases are typically huge and fit almost anything - but this card is a niche product that brings nothing attractive to the table other than its low power consumption and modern encode/decode hardware. It's therefore of far more interest to the HTPC crowd than the gaming crowd, and the HTPC market is chock-full of slim, low-profile, space-constrained, cooling-constrained, incompatible cases that cannot fit your typical gaming GPU these days.

For what it's worth, all is not lost with dumb PCIe power connector locations. Sometimes, you can make an HTPC work using these, assuming the backplate on the card isn't too chunky:

 

[looniam]

So what you're saying is that this 1630 might actually be compliant with the spec if you look at some paywalled info? Fine.

no. i am saying it is compliant to specs to the documentation that i have attained over the years of scouring the internet. i know firsthand google/wikipedia will not give accurate info on what is normally behind a paywall. which is why there is so much misinformation. (education/student subscriptions come in handy!)
paywalls<determination=FTW! all it takes is time.

btw, i have attached them to each post mentioned. so no paywall in your way, so whats problem?

That's great, but it doesn't change a damn thing in the real-world because of that top-edge connector effectively negating the point of the spec even having a maximum height in the first place. Intel and the PCI-SIG can pedantically disagree by a few milimeters as much as they want, but the end result is still the same; The GPU won't fit in someone's case because of where the connector is oriented and you're saying that is omitted from the spec altogether? What a dumpster fire of a spec!

it is all listed in the PCIE-SIG standards, again intel's ATX PSU specs have NOTHING to do with this!

its all there but, even trying to spoon feed you is to no avail

 

[Chrispy_]

Added a test run at 1080p lowest possible settings in all games

Gainward GeForce GTX 1630 Ghost Review - Challenging the AMD RX 6400

The NVIDIA GeForce GTX 1630 launches today. The new card is targeted at the sub-$200 segment and goes head-to-head with AMD's Radeon RX 6400 and Intel's ARC A380. Unlike the Radeon, the GTX 1630 has support for the full PCIe x16 interface.

www.techpowerup.com

Big Oof.
Even at the lowest possible settings it sucks.
Let's assume you have a budget freesync monitor and you can accept 48fps instead instead of 60fps. that's still only a third of the games tested, and that's an AVERAGE not a 99th percentile so you're still going to get dropped frames and stutterfest all over the place :\

no. i am saying it is compliant to specs to the documentation that i have attained over the years of scouring the internet. i know firsthand google/wikipedia will not give accurate info on what is normally behind a paywall. which is why there is so much misinformation. (education/student subscriptions come in handy!)
paywalls<determination=FTW! all it takes is time.

btw, i have attached them to each post mentioned. so no paywall in your way, so whats problem?

it is all listed in the PCIE-SIG standards, again intel's ATX PSU specs have NOTHING to do with this!

its all there but, even trying to spoon feed you is to no avail

Alright captain pedantic, your years of trying to find obscure information have merited you the technical win but I still don't know what point you're trying to make.

My point is crystal clear, I hope; Overly-tall cards, or cards that become tall once you plug in the PCIe connector simply don't fit in many slim cases, which are prevalent in HTPC use.

I'm not sure why you're so obsessed about a 4mm discrepancy between the ATX spec and the PCIE-SIG spec when it doesn't answer any questions that people have asked, or solve any problems that real-world examples encounter.

So, politely, what is your point please?

 

[looniam]

Alright captain pedantic, your years of trying to find obscure information have merited you the technical win but I still don't know what point you're trying to make.

first of all let me remind you of your first post before resorting to name calling,

My point is crystal clear, I hope; Overly-tall cards, or cards that become tall once you plug in the PCIe connector simply don't fit in many slim cases, which are prevalent in HTPC use.

if you would read further, (notes are important!) you would see how that is all addressed. you can have the view the card is too tall for some use cases but saying it doesn't comply with specs is wrong.

I'm not sure why you're so obsessed about a 4mm discrepancy between the ATX spec and the PCIE-SIG spec when it doesn't answer any questions that people have asked, or solve any problems that real-world examples encounter.

i am obsessed with 4mm? refer to above image. and again, ATX specs have nothing to do with this - there are several pages in PCIE-SIG's CEM on brackets and designs to fit in cases and points out what to avoid. and i'll point out that not all case manufacturers adhere to the specs themselves or that the mobo stand off heights are not always appropriate, if you want real world problems.

So, politely, what is your point please?

you're not correct? the info you are relying on is not accurate? i thought that was clear.

 

[ModEl4]

Hi @W1zzard, thanks for the effort!
The 1080p min test, is absolutely minimum in every setting and the AF at 16X on a 5800X, correct?
It gives nearly 2.35X vs ultra regarding average fps score, i wonder when the cross-gen period is going to end, how it will effect the Ultra/min difference (probably will shrink a little bit)

 

[Chrispy_]

first of all let me remind you of your first post before resorting to name calling,

if you would read further, (notes are important!) you would see how that is all addressed. you can have the view the card is too tall for some use cases but saying it doesn't comply with specs is wrong.

i am obsessed with 4mm? refer to above image. and again, ATX specs have nothing to do with this - there are several pages in PCIE-SIG's CEM on brackets and designs to fit in cases and points out what to avoid. and i'll point out that not all case manufacturers adhere to the specs themselves or that the mobo stand off heights are not always appropriate, if you want real world problems.

you're not correct? the info you are relying on is not accurate? i thought that was clear.

I think in this case, "captain pedantic" isn't name calling but an accurate title. If you are offended rather than pleased by that title then you have some reflecting to do because you have been going out of your way to earn it in this thread. Everyone appreciates accuracy and meticulous attention to detail sometimes, but this is not one of those times.

Look, I understand that those 4mm of clearance above the red line are actually within spec, and that I was wrong to use an older, deprecated ATX spec but it doesn't really change anything about the point I'm trying to make, and it doesn't solve the problem for the many of us using slim HTPC cases in the real world. Whether the blame lies with the case manufacturer or the SIG doesn't matter. Trying to get a case that fits your aesthetics under a TV with limited dimensions and cooling in what is typically a short-depth piece of furniture is not easy. We buy what's on the market and if the manufacturer doesn't 100% comply with a specific spec, we can't just send it back and say "hey, please re-tool your entire production line because this card that was released 4 years after your case was released doesn't fit any more".

You still seem hung up on this pedantic irrelevance that a minor deviation in height is the issue - and it has nothing to do with my point. Real world height is what it is, including the airgap required for the heatsink to exhaust and the additional height any mandatory cables require. We're already off-topic enough and bickering about minor differences in spec really isn't adding anything to this review.

 

[W1zzard]

The 1080p min test, is absolutely minimum in every setting and the AF at 16X on a 5800X, correct?

absolutely minimum everything including AF. So AF set to trilinear where available. Exact same test system

 

[looniam]

I think in this case, Captain pedantic isn't name calling but an accurate title. If you are offended rather than pleased by that title then you have some reflecting to do because you have been going out of your way to earn it in this thread. Everyone appreciates accuracy and meticulous attention to detail sometimes, but this is not one of those times.

you do not speak for everyone, i am more than sure the peanut gallery appreciates getting docs that are not publicly accessible and how to get them.

Look, I understand that those 4mm of clearance above the red line are actually within spec, and that I was wrong to use an ATX spec

gee would that have been so hard to realize in the beginning?

but it doesn't really change anything about the point I'm trying to make, and it doesn't solve the problem for the many of us using HTPC cases in the real world.

You are hung up on this pedantic irrelevance about a 4mm deviation between the spec I found and the actual current 2022 spec, not me, and it has nothing to do with my point.

look pal, i pointed everything out in my first reply; this whole conversation since then is also on you. if you really like titles, i like to hand out knucklehead from time to time.

and btw, no i cannot find the newest CEM since what i posted, so if you have it, then by all mean share.

until then nice chat.

 

[Chrispy_]

look pal, i pointed everything out in my first reply; this whole conversation since then is also on you. if you really like titles, i like to hand out knucklehead from time to time.

LOL what? My first response you was this:

TBH the first result Google returned was ATX 2.2 and I didn't spend long hunting, I just hit CTRL+F

I'm saying right there, at the beginning of this entertaining discussion that I didn't really look too hard and admitting it's an old spec. I'm not pretending that I care or that I know better. If that makes me a knucklehead according to you then I can live with that.

Out of interest, what's a CEM? I'm 99.9% certain I don't have the newest one of them.

 

[catulitechup]

Added a test run at 1080p lowest possible settings in all games

Gainward GeForce GTX 1630 Ghost Review - Challenging the AMD RX 6400

The NVIDIA GeForce GTX 1630 launches today. The new card is targeted at the sub-$200 segment and goes head-to-head with AMD's Radeon RX 6400 and Intel's ARC A380. Unlike the Radeon, the GTX 1630 has support for the full PCIe x16 interface.

www.techpowerup.com

@W1zzard

thanks for your results but this only reconfirm 1080p for this card and other similar are too much, 720p test will be more interesting and more usable than 1080p

 

[W1zzard]

720p

Are people actually playing games at 720p? try it, it looks terrible

 

[looniam]

LOL what? My first response you was this:

I'm saying right there, at the beginning of this entertaining discussion that I didn't really look too hard and admitting it's an old spec. I'm not pretending that I care or that I know better. If that makes me a knucklehead according to you then I can live with that.

well i don't really get into he said/he said narratives of what is already posted. just reread the comments for any questions.

Out of interest, what's a CEM? I'm 99.9% certain I don't have the newest one of them.

Card Electromecanical Specfications that are listed in the newest ATX 3.0 guide:

maybe read it?

thats the problem, you had argued as if you did have them.

glad we got that cleared up have a good day.

 

[Ruined Mind]

Are people actually playing games at 720p? try it, it looks terrible

I always use either the 1280 x 720 resolution or the 1360 x 768 resolution. My screen's resolution is 1360 x 768. Using the 1280 x 720 resolution and enlarging it to fill the 1360 x 768 canvas looks fine to me.

 

[Chrispy_]

Are people actually playing games at 720p? try it, it looks terrible

Just about acceptable on a 13" laptop, which is also the only class of device likely to be running solely UHD Graphics 620 with a single-channel of soldered DDR4-2400.
It's not pretty, but sometimes it's necessary.

Thankfully most games have resolution scaling these days which at least renders the HUD and UI at native resolution. FSR/DLSS are better but if you have to run at 720p your hardware probably isn't powerful enough to use either of those options.

 

[AusWolf]

Hello. I need a low-profile dual-slot card. (Gigabyte has a low-profile version of the 1630.)

I'm afraid to buy a used card, because of the lack of a warranty. Even if I'd consider a used one, I won't buy a 1650, because I've noticed reviews about the low-profile versions of the 1650 from people who said the fans were so loud that they decided to return them and stay with their much quieter 1050 Ti cards. So, if I'd consider a used one, I'd choose the 1050 Ti. I have a PCI-Express 2.0 slot. Please don't tell me to buy a new computer. My i5-2400 processor hasn't been a bottleneck in the games I've played. I've checked with MSI Afterburner.

Regardless of whether I'd buy a used one, I still want an answer to this question from the professionals of this website, especially the leader called "W1zzard":
Would an RX 6400 be so limited in a PCI-Express 2.0 x4 configuration that the 1630 would be faster, or would the RX 6400 still be faster than the 1630, even in a nasty 2.0 x4 situation?

Hi there.

Low profile 1650 variants can be good or bad. I used to have an Asus one that was great. I still regret selling it to this day. I've heard Gigabyte and MSI are crap, but can't confirm it.

As for the 6400 in a pci-e 2.0 board, its performance will hugely depend on the game you wish to play. Some of them will suffer no performance hit, some of them might be crippled badly. If you let me know what you want to play, and if I happen to have it, I'll gladly test it for you and PM you the results.

 

[catulitechup]

Are people actually playing games at 720p? try it, it looks terrible

yeah many people still using 720p (old machines where supposed gtx 1630 offer improvement) and more on too weak gpus like gtx 1630 and similars

1080p is too much for this hardware

 

[Athlonite]

@Chrispy_ & @looniam if you two want to argue take it to your DM's otherwise behave like grown ups and stick to being on topic

 

[Mussels]

On the AV1 and video streaming

Literally, on PC onlyyoutube lets you do 4K content easily.
Netflix requires an add-on HEVC codec to enable 4K, that has its own conditions and limitations - and they will sooner or later do the same with AV1
Disney+ is still locked at 720p even in the windows store app

It's a nightmare situation where modern CPU's block you from playing bluray, and modern GPU's can block you from playing streaming content in 4K, so anything that makes that situation more complex is just not fun.

(intel 6th gen through 10th gen work for 4k blu ray and 4k streaming, but nothing older or newer can do both)

 

[the ram]

It's possible flash the BIOS with a gtx 1650 sample? Or something?

 

[Udyr]

Are people actually playing games at 720p? try it, it looks terrible

That +/- 12% under 1080p should greatly benefit from a product like this, at the right price. Unfortunately it's a no-go for them at the moment.