The headline additions fall into four distinct categories, and each one tells a different story about where the graphics industry is headed in the second half of 2026. NVIDIA dominates the changelog with consumer, professional, and region-specific variants; AMD fills in budget gaps that have been empty for months; Intel continues its quiet embedded push; and Moore Threads signals that China’s domestic GPU effort hasn’t stalled despite U.S. export restrictions tightening through 2025.

RTX 5090 D V2 Confirms a Second China-Market Revision

The most eyebrow-raising entry in GPU-Z 2.69’s support list is the GeForce RTX 5090 D V2 — a designation that confirms NVIDIA has already revised its China-export variant of the flagship Blackwell consumer GPU. The original RTX 5090 D shipped as a compliance-trimmed version designed to meet U.S. export control thresholds, with reduced CUDA core counts and memory bandwidth caps compared to the unrestricted RTX 5090 sold in other markets. A “V2” revision this quickly after launch typically means one of two things: either NVIDIA adjusted the hardware to comply with updated export rules from the Bureau of Industry and Security, or the company optimized the silicon binning process to improve yields on the restricted SKU.

Having tracked NVIDIA’s China-market GPU variants since the original RTX 4090 D appeared in late 2023, the V2 pattern feels more regulatory than commercial. Washington tightened compute density thresholds twice during 2025, and each adjustment forced NVIDIA to re-spec its China offerings — a rolling compliance headache that GPU-Z’s database quietly documents with each new entry. The RTX 3050 A Mobile also appearing in this update follows the same logic: a region- or OEM-specific variant that exists because global GPU distribution has become a regulatory maze rather than a simple supply chain.

NVIDIA’s Full RTX Pro Blackwell Workstation Stack Revealed

GPU-Z 2.69 adds support for eight RTX Pro Blackwell variants, and the sheer breadth of the lineup — from the RTX Pro Blackwell 5000 down to the RTX Pro Blackwell 500 — confirms that NVIDIA’s professional GPU rebrand is far more extensive than the company has publicly detailed. The naming scheme abandons the old Quadro-era numbering (RTX A4000, A5000) in favor of a tiered approach that mirrors the consumer GeForce stack: 5000 for flagship workstation use, 4500/4000 for mid-range CAD and visualization, 2000/1000 for entry-level professional certification, and 500 for basic multi-display office configurations.

What caught my attention is the inclusion of both SFF and Embedded variants alongside standard desktop and mobile SKUs. The RTX Pro Blackwell 4000 SFF targets small-form-factor workstations — a growing segment driven by studios and engineering firms replacing tower machines with compact nodes — while the RTX Pro Blackwell 5000 Embedded aims at industrial and medical imaging applications where GPU compute runs inside sealed, fanless enclosures. NVIDIA hasn’t published official specs for most of these models, so GPU-Z’s device ID entries represent the first concrete evidence that the full product matrix is locked and silicon is reaching validation stages. For IT departments planning workstation refresh cycles, this list is more useful than anything NVIDIA’s professional sales team has shared publicly.

Moore Threads S30 and the Quiet Expansion of Chinese GPUs

The Moore Threads S30 addition deserves more attention than it’ll get in most English-language coverage. Founded in 2020 by former NVIDIA China chief Zhang Jianzhong, Moore Threads has been building x86-compatible GPUs designed specifically for the Chinese domestic market — targeting government, enterprise, and data center buyers who need alternatives to restricted NVIDIA and AMD hardware. The S30 appears to be a successor or variant of the company’s MTT S80 series, which offered roughly GTX 1060-class performance when it launched in 2022 alongside a custom MUSA architecture that intentionally avoided dependencies on Western IP.

GPU-Z adding Moore Threads support signals that these GPUs are reaching a maturity level where standard Western diagnostic tools need to account for them — either because evaluation units are circulating outside China, or because the installed base has grown large enough that TechPowerUp’s user telemetry flagged the gap. Neither scenario is trivial. Four years ago, the idea that a Chinese-designed GPU would appear in GPU-Z’s database alongside GeForce and Radeon entries would have seemed premature; today it reflects a supply chain bifurcation that’s reshaping the global semiconductor landscape in real time.

AMD and Intel Round Out the Budget and Embedded Tiers

On the AMD side, GPU-Z 2.69 adds support for the Radeon RX 7400, Radeon 820M, Radeon Pro W6600X, Radeon Pro W7800, and Radeon Pro V710. The RX 7400 fills a gap that AMD’s consumer lineup has desperately needed — a true sub-$200 RDNA 3 card for budget builders who can’t justify the RX 7600’s $250+ street price. The Radeon Pro additions suggest AMD continues pushing into the professional visualization market, though the W6600X’s appearance (based on older RDNA 2 architecture) indicates that some workstation OEMs still need driver-certified options for legacy compute workflows.

Intel’s entries — the Arc A380E and Arc A310E — carry the “E” suffix that designates embedded and industrial variants, joining Intel’s quiet push into digital signage, kiosk, and edge computing applications where discrete GPU capability matters but gaming performance doesn’t. These aren’t the Arc GPUs that Intel wants headlines about, but they represent steady revenue from design wins that don’t depend on consumer sentiment or Steam hardware surveys.

Beyond new GPU support, version 2.69 fixes a shader count calculation bug affecting certain AMD and Intel GPUs — a correction that matters for anyone using GPU-Z to verify hardware specifications before purchasing used cards, where misreported shader counts could mask a partially disabled die. Updated device IDs for NVIDIA’s older Quadro P1000, Quadro P2000, and RTX 6000 round out the maintenance work, along

Sources: TechPowerUp GPU-Z, WCCFtech

The post GPU-Z 2.69 Drops Support for RTX 5090 D V2, Pro Blackwell, and China’s Moore Threads S30 appeared first on Geeknify.

Following the retirement of Paint 3D, Microsoft has now confirmed plans to end support for 3D Viewer, a built-in utility that originally shipped with Windows 10. The company says the app will be removed from the Microsoft Store on July 1, 2026, effectively marking the end of its lifecycle.

3D Viewer has already been optional in Windows 11, but this change makes its status permanent. Users who currently have the app installed will still be able to run it, but once it is manually removed, there will be no official way to download or reinstall it.

Microsoft originally positioned 3D Viewer as a lightweight tool for viewing and making basic edits to 3D models in formats such as STL, FBX, OBJ, and GLTF. While it was never intended for professional workflows, it proved useful for quick previews and simple inspections without requiring full-featured third-party software.

As an alternative, Microsoft is directing users to Babylon.js Sandbox, a browser-based viewer that runs entirely on the web. The move reflects a broader shift in Microsoft’s Windows strategy, where web-based tools increasingly replace native utilities for niche or low-usage tasks.

For long-time Windows users, the removal of 3D Viewer follows a familiar pattern. Built-in apps are gradually being phased out as Windows evolves, often without direct replacements at the OS level, leaving users to rely on external or web-based solutions instead.

The post Microsoft is retiring 3D Viewer and removing it from Windows in 2026 appeared first on Geeknify.

The first addition is an advanced frame generation mode capable of producing up to six AI-generated frames for every traditionally rendered frame. The second is a new Dynamic Multi Frame Generator, which takes a more flexible approach to performance smoothing. Instead of relying on a fixed frame-generation multiplier, it lets users set a target frame rate. DLSS then dynamically adjusts how many generated frames are inserted to hold that target as performance rises or falls.

For anyone who has used frame generation in modern games, the appeal is obvious. Performance rarely stays consistent — busy combat scenes, heavy effects, or large open areas can cause sharp swings in frame rate. By reacting in real time, Nvidia’s dynamic approach aims to smooth out those drops without forcing users to guess which multiplier will work best.

There is a clear limitation, however. Both features are expected to launch in April, meaning players will have to wait a bit longer before testing them firsthand. As with Nvidia’s newest frame-generation technologies, these capabilities will also be exclusive to GeForce RTX 50-series GPUs.

While DLSS 4.5 itself already works across all GeForce RTX cards, Nvidia is drawing a firm line between current support and what it considers next-generation frame generation. More aggressive multi-frame techniques and dynamic controls are being positioned as features that rely on the additional AI and hardware capabilities of the RTX 50 series.

For now, RTX 40- and 30-series owners can use DLSS 4.5 as it stands, but the most ambitious frame-generation advances will remain locked to Nvidia’s latest GPUs when they arrive later this spring. For a deeper look at how it stacks up against AMD’s alternative, see our full Nvidia DLSS 4.5 and AMD FSR 4 comparison in modern games.

Sources: HardwareLuxx, Videocardz

The post Nvidia’s next DLSS features will push up to Six AI frames per real one appeared first on Geeknify.

One of the most visible cases involves the Radeon 9070 XT, where undervolting settings that worked flawlessly on earlier drivers now cause immediate instability.

Stable undervolts no longer work

A Reddit user going by SnooJokes5264 shared their experience, noting that an undervolt of –50 mV combined with a –9% power limit was completely stable on Adrenalin 25.12.1. After updating to 26.1.1, however, those same settings began causing instant game crashes.

Launching Cyberpunk 2077 with the old profile results in a crash at startup. Stability only returns after resetting all tuning options to factory defaults, suggesting the issue is directly tied to manual voltage and power adjustments rather than the game itself.

Possible changes under the hood

While AMD has not officially acknowledged the problem, the behavior points to silent changes in voltage curves or frequency control algorithms for newer GPU architectures in driver version 26.1.1.

If that’s the case, it would explain why undervolting values that were once safe now push the GPU beyond its new stability thresholds. Even small offsets that previously reduced temperatures and power draw without side effects may now trigger crashes.

Adrenalin 26.1.1 appears to alter how newer Radeon GPUs handle voltage and power, effectively invalidating previously stable undervolt profiles. For now, users focused on efficiency rather than stock performance may want to hold off on updating or be prepared to retune from scratch.

As always with GPU drivers, stability beats small gains, especially when the changes happen without warning.

Source: Reddit

The post AMD Adrenalin 26.1.1 reportedly breaks undervolting on new GPUs appeared first on Geeknify.

Logic says the newest OS should win. Cleaner code, smarter schedulers, better optimization. That assumption collapses the moment you look at the actual results. In this test, Windows XP – released in 2001 and now mostly found in ATMs – regularly outperformed Windows 11.

It sounds wrong. It isn’t.

The test setup: Same hardware, six versions of Windows

YouTube creator TrigrZolt ran the experiment using six identical Lenovo ThinkPad X220 laptops, each powered by an Intel Core i5-2520M (Sandy Bridge), 8 GB of RAM, and a 256 GB hard drive. Each system received one Windows version, fully updated:

• Windows XP
• Windows Vista
• Windows 7
• Windows 8.1
• Windows 10
• Windows 11

This hardware was never meant for Windows 11. There’s no TPM 2.0, the CPU is from 2011, storage is a spinning HDD, and UEFI support is… generous at best. The OS had to be installed using workarounds.

That’s exactly why the test is interesting. When resources are limited, inefficiencies stop hiding.

Boot time: Windows 11 loads fast

Boot speed was the first metric tested. Surprisingly, Windows 8.1 came out on top thanks to Fast Boot and a trimmed startup sequence. Vista, long mocked for sluggishness, placed second. Windows XP took third.

Windows 11 finished dead last.

The reason isn’t the kernel. It’s the interface.

Once the core system loads, Windows 11 waits for the taskbar to initialize. Unlike older versions where the taskbar lived inside Explorer.exe, Windows 11’s taskbar is built on WinRT and WinUI 3. That means additional runtimes, services, and UI frameworks must initialize before the desktop is usable.

Fast Boot still exists, but its gains are eaten by slow UI initialization.

Disk space: Modern Windows is heavy by design

Installed disk usage follows a simple rule: newer Windows versions are larger because they rarely remove old components.

• Windows XP — 6.5 GB
• Windows Vista — 15.3 GB
• Windows 7 — 17.4 GB
• Windows 8.1 — ~18 GB
• Windows 10 — ~25 GB
• Windows 11 — 29.8 GB

Backward compatibility comes at a cost. Libraries stick around to prevent legacy apps from breaking, and over time the OS grows bloated. On budget laptops with small, non-upgradable drives, this becomes a real limitation.

RAM usage: Where Windows 11 starts losing the fight

Idle memory usage paints an even clearer picture:

• Windows XP — 0.8 GB
• Windows Vista — ~1.5 GB
• Windows 7 — ~1.8 GB
• Windows 8.1 — ~1.9 GB
• Windows 10 — 2.0 GB
• Windows 11 — 3.3–3.7 GB

Before opening a single app, Windows 11 is already consuming nearly half of an 8 GB system. The reasons are familiar:

• background accessibility services
• file indexing
• Microsoft Store updates
• WebView2 for system apps
• widgets
• Teams integration
• OneDrive sync
• telemetry and diagnostics

On systems with limited RAM, Windows 11 quickly spills into the paging file. On a mechanical hard drive, that’s where responsiveness goes to die.

Browser stress test: Tabs tell the truth

To measure real-world multitasking, the test pushed each OS until total memory usage hit 5 GB by opening browser tabs.

• Windows 8.1 — 252 tabs
• Windows 7 — over 200 tabs
• Windows Vista — over 100 tabs
• Windows 10 — over 100 tabs
• Windows XP — 50 tabs (paging crashes beyond that)
• Windows 11 — 49 tabs

Despite being ancient, Windows XP still matched Windows 11 — simply because it doesn’t waste gigabytes on background services.

App performance: When lightweight wins

Audacity was used to export the same audio file across all systems. Results ranked best to worst:

1. Windows 8.1
2. Windows XP
3. Windows 7
4. Windows 10
5. Windows 11
6. Windows Vista

Windows 11’s heavier audio stack introduces more latency. Where XP used a short path to the driver, modern Windows routes audio through multiple permission and processing layers. On slow storage, every extra step hurts.

Launching everyday apps like File Explorer, Paint, Calculator, Adobe Reader, and VLC told a similar story. Windows 11 consistently finished last, with Windows 10 taking first place.

Even File Explorer struggled. Its WinUI 3 rewrite, OneDrive hooks, and widget integration roughly doubled memory usage compared to Windows 10’s version.

Benchmarks vs reality

Synthetic benchmarks didn’t save Windows 11 either.

In CPU-Z single-core testing, Windows XP actually took first place, narrowly beating Windows 7. Windows 11 landed fourth. Multi-core results showed the same trend.

Geekbench 6 favored Vista, while Windows 11 edged out Windows 10 in single-core but lost in multi-core workloads. The scheduler is optimized for modern hybrid CPUs, not an old dual-core Sandy Bridge chip – and that mismatch costs performance.

Windows 11 feels slower than XP

This isn’t about bad coding. It’s about architecture.

Older Windows versions relied on GDI and User32, drawing UI through a short, efficient path. Modern Windows uses WinUI 3, WinRT, DirectComposition, and GPU-driven effects. That enables smooth animations, transparency, and blur – but only if the hardware can keep up.

Security adds more overhead. Windows Defender now intercepts nearly every file operation. Telemetry services constantly wake up. On SSDs, this is barely noticeable. On HDDs, it’s brutal.

Why Windows 8.1 won the test

Windows 8.1 accidentally hit the sweet spot.

It introduced Fast Boot, better memory management, and an improved scheduler without dragging along modern UI bloat. It was widely disliked for its interface, but under the hood it remained lean.

Fast Boot works best when the OS itself is lightweight. Windows 8.1 still is. Windows 11 isn’t.

Is 8 GB of RAM enough for Windows 11?

On modern hardware with NVMe storage and 16–32 GB of RAM, Windows 11 behaves very differently. This test used hardware it was never designed for.

Still, the trend is clear: Windows no longer gets lighter with age. Each generation adds features, services, and visual polish. That’s great for new machines. It’s punishing for older ones.

For users on aging laptops or minimal configurations, Windows 11 simply isn’t efficient. Progress has a cost.

The only real question is who ends up paying it.

Source: TrigrZolt (YouTube)

The post Why Windows XP can outrun Windows 11 on identical hardware appeared first on Geeknify.

The update is rolling out on macOS, Windows, and Chromebook Plus, with some features initially limited to users in the United States.

Gemini moves into Chrome’s sidebar

The most visible change is where Gemini now lives. Instead of opening in a separate window, it’s embedded directly into Chrome’s sidebar, staying accessible over any open tab.

That small shift makes a big difference. You can keep reading an article, browsing a store, or checking flight options while simultaneously asking questions, comparing information across sites, or pulling quick summaries – without losing your place or breaking your flow.

It’s especially useful for research-heavy tasks like comparing product reviews, planning trips, or scanning long pages for key points.

Image editing and visual tools, right in the browser

Chrome now supports an image tool called Nano Banana, letting users work with visuals directly on the page.

You can edit images without downloading files or opening third-party apps. It’s designed for practical use cases – mocking up interior design ideas, creating simple visuals for presentations, or turning online data into quick infographics.

Everything happens inside the browser, which keeps the process fast and surprisingly friction-free.

Deep integration with Google Services

Gemini in Chrome connects tightly with Google’s own ecosystem. This includes:

• Gmail
• Google Calendar
• YouTube
• Google Maps
• Google Flights
• Google Shopping

With permission, the assistant can pull context from emails, calendar events, or saved plans. That means it can help draft messages, surface relevant travel details, suggest routes, or find products based on what you’re already working on.

All integrations can be managed and limited through Gemini’s settings, giving users control over what’s connected.

Auto Browse: Chrome that can do the clicking for you

For U.S. users subscribed to AI Pro and Ultra plans, Google is introducing a new feature called Auto Browse.

This turns Chrome into something closer to a task-runner than a traditional browser. Auto Browse can:

• search for hotels and flights
• fill out forms
• gather documents
• review bills and subscriptions
• handle routine online tasks

It also supports visual input. Show it an image, and it can identify similar products, compare prices, apply discounts, add items to a cart, and stay within a budget you define.

Purchases and other sensitive actions always pause for user confirmation before anything is finalized.

Built with safety in mind

Google emphasizes that these features were designed with safeguards from the start. Any action involving payments, public posting, or account changes requires explicit approval.

Automatic workflows stop when confirmation is needed, ensuring users stay in control even when Chrome is doing most of the heavy lifting.

The first step toward a more active WEB

Google describes this update as an early move toward an “agent-driven web.” In practice, that means browsers stop acting like passive viewers and start functioning as helpers saving time on repetitive tasks and reducing friction across everyday online activities.

The post Google is turning Chrome into more than a browser appeared first on Geeknify.

Instead, Microsoft spent the opening weeks of 2026 dealing with a messy chain of problems that cut across Windows, the cloud, and user trust. A Windows 11 update caused unexpected system failures. Cloud services went dark for hours. And a separate privacy issue reminded users that encryption inside a Microsoft account comes with conditions.

Each incident on its own would’ve been manageable. Together, they felt harder to dismiss.

Windows 11 Updates that tried to do too much

On January 13, Microsoft pushed out a cumulative Windows 11 security update, KB5074109. Along with the usual fixes, it brought new features – most notably a gaming-oriented Full Screen Experience mode.

That ambition didn’t age well. Within days, users started reporting problems that were hard to ignore: apps crashing for no clear reason, desktop wallpapers turning into solid black screens, Outlook freezing mid-session, and systems failing to properly enter sleep or hibernation. This wasn’t subtle breakage. It was the kind of stuff that forces people to stop what they’re doing.

Microsoft acknowledged the issues four days later and released an out-of-band patch. It helped in some cases, but the reports didn’t stop. Apps tied to cloud sync – OneDrive especially – kept crashing or hanging. At that point, Microsoft’s advice was telling: move important files to local storage, at least for now. A second emergency patch followed soon after.

When a fix becomes the bigger problem

That second update, KB5078127, bundled earlier fixes and aimed to stabilize cloud-related behavior. For most systems, it did. For a smaller group, it created a much bigger headache.

Some machines failed to boot at all, landing users on a black screen with the error UNMOUNTABLE_BOOT_VOLUME – usually a sign of storage or firmware trouble. Microsoft says the issue affects a limited number of devices and recommends recovery through WinRE. Still, when a security update risks locking people out of their systems, confidence drops fast.

BitLocker, defaults, and an uncomfortable reminder

While Windows 11 issues were still unfolding, Microsoft confirmed something else that caught users’ attention. The company acknowledged providing BitLocker recovery keys to the FBI under a court order, allowing access to encrypted data on a Windows device.

Legally, this wasn’t shocking. Microsoft says it handles roughly 20 such requests each year. What unsettled people was the reminder of how BitLocker works by default. Recovery keys are often tied to a Microsoft account rather than stored only on the device itself. It’s convenient when things go wrong – but it also means Microsoft can supply those keys when legally required.

Users can change that behavior and store keys locally. Many never do, often without realizing there’s a choice.

Privacy advocates argue this blurs the line between user-controlled encryption and platform-managed access. And once that mechanism exists, it doesn’t stop at U.S. borders.

When the cloud stops cooperating

Later in the month, Microsoft ran into another problem. Between January 22 and 23, a large-scale outage knocked out access to Outlook, OneDrive, Teams, the Microsoft Store, Copilot, and related services. It took close to ten hours to fully restore everything.

Microsoft blamed a spike in demand combined with temporary capacity limits during scheduled maintenance in North America. The explanation was plausible. The experience wasn’t reassuring.

For companies and users deeply embedded in Microsoft’s ecosystem, the message was simple: when the cloud stalls, work stalls with it.

What January quietly exposed

Microsoft will fix the bugs. Cloud reliability will improve. Legal obligations aren’t going away. The more interesting takeaway from January is how tightly everything is now connected – and how little slack there is when something breaks.

Windows isn’t just an operating system anymore. It’s an entry point to identity systems, cloud storage, productivity tools, and encrypted data. When stability, availability, and trust all wobble at the same time, people notice.

January wasn’t a disaster. But it was revealing.

The post Microsoft’s Black January was more than a bad update appeared first on Geeknify.

Clawdbot combines agent-style AI behavior, extensive third-party integrations, and an open-source foundation. More importantly, it doesn’t rely solely on centralized cloud services. Users can run it locally on their own computer or deploy it on a private cloud server they control.

That alone puts it in a different category from most mainstream assistants.

How Clawdbot works

Clawdbot runs quietly in the background and connects to a chat provider of the user’s choice. Supported options include:

• ChatGPT
• Claude
• DeepSeek
• Google-based models
• Locally hosted models

The assistant continuously builds context over time, learning from previous interactions instead of treating every request as isolated.

This makes it feel less like a chatbot – and more like a persistent digital assistant.

Chat аirst, platform agnostic

Instead of forcing users into a proprietary interface, Clawdbot works through platforms people already use every day.

You can interact with it via:

• Telegram
• WhatsApp
• Discord
• Slack
• Signal
• iMessage

From a usability standpoint, this is one of its strongest ideas. There’s no new app to learn – just a familiar chat window.

Deep system and service integrations

Where Clawdbot really stands out is integration depth.

It can:

• Read, create, and modify local files
• Interact with the operating system
• Control a web browser
• Run automated workflows
• Respond to voice commands

On top of that, it connects to popular services such as:

• Spotify
• X (Twitter)
• GitHub
• Notion
• Gmail
• Home Assistant

This turns Clawdbot into a control layer between the user, their apps, and their system.

Real-world use cases

Early adopters are already finding practical ways to use it:

• Generating daily email summaries
• Detecting scheduling conflicts
• Automating routine work tasks
• Ordering food through connected services
• Managing smart home devices
• Even replying on a user’s behalf in group chats

That last use case has sparked some debate – but it also highlights how flexible the system can be.

Local control and privacy appeal

One reason Clawdbot resonates with technically minded users is control.

Because it can run locally or on a privately rented server:

• Data doesn’t have to live on third-party platforms
• Users decide which services it can access
• Integrations can be enabled or disabled granularly

For developers and power users, this is a major shift away from black-box assistants.

Getting started

The project provides installation and setup documentation in English on its official website. While initial configuration requires some technical comfort, the process is well documented and clearly aimed at users who value customization over simplicity.

This isn’t a plug-and-play assistant for everyone – and that’s very much the point.

Final thoughts

Clawdbot isn’t trying to compete with consumer-focused assistants built into phones or operating systems. Instead, it targets users who want flexibility, transparency, and control.

If the project continues to mature, it could become a serious option for anyone looking to build a personalized AI workflow that actually fits their daily tools – not the other way around.
Setup and installation guides for Clawdbot are available on the official project website.

The post Clawdbot is gaining attention as a Local, Open-Source AI Assistant with deep integrations appeared first on Geeknify.

Here’s a clear breakdown drawn from recent tests and expert evaluations, with practical insights for gamers who care about both visuals and smooth play.

What is DLSS 4.5 — and why it matters

DLSS has evolved from a simple AI upscaler into a multi-faceted performance and visual enhancement tool. DLSS 4.5 builds on the transformer-based models introduced in earlier versions by:

• Upgrading Super Resolution with a second-generation transformer for sharper, more stable images.
• Introducing new presets like Model M (Performance) and Model L (Ultra Performance).
• Preparing the groundwork for Dynamic Multi Frame Generation and 6× frame generation on newer RTX cards.

For RTX 4070 gamers, the appeal is straightforward: better visuals with minimal compromise — if your setup can handle it.

RTX 4070 performance in Cyberpunk 2077 — What early benchmarks show

DLSS 4.5’s impact can vary based on resolution, settings, and GPU generation. While specific 4070 Cyberpunk benchmarks are still rolling out, community data and early tests paint a consistent picture:

Visual quality gains

• DLSS 4.5 generally improves image sharpness, reduces shimmering artifacts, and enhances fine detail compared to DLSS 4.
• Preset L (Ultra Performance) often balances upscaling quality and performance at higher resolutions.

Performance Trade-offs

• On older RTX cards (30 series), users have seen notable FPS drops (14–24%) with DLSS 4.5 vs DLSS 4.
• The RTX 4070 should fare better than older GPUs, thanks to FP8 acceleration supporting heavier compute loads, but mild performance cost is still possible — especially at 4K.

Real-world implication

In practical terms: in demanding titles like Cyberpunk 2077, enabling DLSS 4.5 often yields smoother visuals than older DLSS modes, but you may see slightly lower average FPS than with DLSS 4. Whether that’s “worth it” depends on how much you value image quality over raw frame rate.

Settings that make sense for RTX 4070 players

Here’s a starting point if you’re tweaking Cyberpunk 2077 on a 4070:

For 1440p Gaming

• Use DLSS 4.5 Model M for a balanced experience — good visuals with solid performance.
• Avoid putting ray tracing at maximum unless frame rate dips below your target (e.g., 60 FPS).

For 4K gameplay

• Model L (Ultra Performance) is usually the better DLSS 4.5 choice at native 4K to keep frame rates playable.
• Consider lowering less noticed post-process effects (such as shadows and crowd density) to offset GPU load.

These recommendations blend community insights with common best practices, but don’t be afraid to experiment — Cyberpunk reacts differently to upscaling vs raster settings.

Why some players hesitate

Not every RTX owner will rush to DLSS 4.5 immediately. A few realities you should know:

• Higher computational cost: DLSS 4.5 is more demanding than earlier versions, and older GPUs will feel it more.
• Variable gains by game: Improvements depend on how a title implements upscaling and frame generation. Some scenes benefit more than others.
• Driver & support matters: To access the newest DLSS 4.5 features, you’ll need the latest NVIDIA app update and drivers.

Bottom line: Should RTX 4070 owners use DLSS 4.5?

If you play Cyberpunk 2077 on a 4070, DLSS 4.5 is definitely worth testing — especially at 1440p or 4K where visual fidelity gains are most apparent.

You might trade a few frames compared to DLSS 4, but many users will find the visual improvements and smoother upscaled imagery a justifiable exchange. Just bear in mind this isn’t an across-the-board performance miracle: results will differ by map, ray tracing load, and resolution.

Source: youtube

The post NVIDIA RTX 4070 & DLSS 4.5 in Cyberpunk 2077: Real-World performance deep dive appeared first on Geeknify.

Both upscalers represent the latest advances from their respective developers. DLSS 4.5 builds on Nvidia’s transformer‑based super sampling and multi‑frame generation tech to offer improved image reconstruction and dynamic performance scaling. Meanwhile, FSR 4 from AMD brings enhanced spatial upscaling and frame generation features, expanding its compatibility across titles supported by the technology.

Cyberpunk 20277 Benchmarks – DLSS 4.5 & FSR4:

Performance and image quality differences

In the side‑by‑side comparisons, both upscalers showed clear improvements over previous versions and native rendering at high resolutions. Tests highlighted that DLSS 4.5 often delivered higher overall image clarity and stability, especially in motion, and tended to preserve detail better in complex scenes. Many gamers and reviewers pointed to DLSS 4.5’s advanced transformer model and multi‑frame generation as giving it an edge in visual fidelity and responsiveness.

That said, FSR 4 has made noticeable strides compared with older upscaling tech and narrowed the gap with Nvidia’s solution in many scenarios. In titles that support it, FSR 4 provided competitive visuals and performance, although some side‑by‑side comparisons suggested the overall quality and temporal stability still leaned in Nvidia’s favor.

Community feedback also reflects this mixed landscape: some users report that DLSS 4.5 feels visibly sharper and cleaner at the same performance tier, while others note that FSR 4 is “close enough” in most cases and offers broader compatibility on AMD hardware.

Hogwarts Legacy Benchmarks – DLSS 4.5 & FSR4:

Real‑world context and how it matters

For PC gamers, the practical difference between these technologies can affect not only visuals but also framerate targets and smoothness. DLSS 4.5’s dynamic multi‑frame generation often helps Nvidia GPUs push higher average frame rates at 4K and high refresh rates, which is especially noticeable on RTX 40 and 50 series cards. FSR 4’s open and hardware‑agnostic nature means it works across a broader range of devices — including Radeon cards — but may not achieve the same level of detail reconstruction in every title.

Ultimately, both upscalers are effective tools for modern gaming. DLSS 4.5 tends to shine in image quality and performance scaling, while FSR 4 continues to evolve and offers solid support where Nvidia’s tech isn’t available. Gamers choosing between them should consider game support, hardware compatibility, and personal preferences for visual quality versus raw performance.

Sources: nvidia, mobidevices, wccftech

The post Nvidia DLSS 4.5 and AMD FSR 4 compared in modern games appeared first on Geeknify.