1998 research cohort / GPU
ATI · model variant
L4 · reviewed discrepancyATI RAGE 128GL graphics chip is recorded in the 1998 GPU research cohort under ATI. The current display name and classification are normalization proposals.
ATI addresses workstation/high-end 3-D graphics
Intended users: workstation and high-end OEM graphics systems
Engineering response: RAGE 128GL 128-bit graphics chip with OEM sampling and planned production
ATI announces RAGE 128 family · lines 7-18
Source distinguishes GL as workstation/high-end and reports sampling/production plans.
Plans conflict with MPR and do not prove shipment; no independent benchmark retained.
Comparable benchmark evidence has not been established for this record. Missing results do not imply poor performance.
The year above is a research-cohort label. Announcement, production, shipment and retail availability are distinct events.
Chip identity is stable, but forecast timing conflicts and does not establish actual shipment.
PC Watch / Impress · lines 7-18
RAGE 128GL is the 128-bit workstation/high-end chip; announcement, sampling, and September plan.
Keep the distinct GL chip record and preserve the conflicting September and October production forecasts; neither forecast proves shipment.
What would resolve it: Dated evidence establishing the actual production or commercial shipment event and resolving the September-versus-October statements.
Actual shipment, plan outcome, chip/board SKU, OEM/region.
No duplicate: GL and VR are separately named RAGE 128 variants with different interface/package descriptions in the registered contemporary reporting.
Audit source · article lines 5-17
The report separately identifies RAGE 128GL and RAGE 128VR and describes distinct 128-bit and 64-bit configurations.
Follow-up: Retain variant-level rows; do not merge without an explicit family/variant policy decision.
Current rows are variants, not an independently evidenced family launch row; a parent-family relation may be useful but is not a new candidate in this audit.
Audit source · article lines 5-17
The source names both variants in one RAGE 128 family context while giving separate package/interface details.
Follow-up: Record a non-canonical parent relation during normalization; do not add a family candidate from this finding.
Planned October board availability cannot be promoted to confirmed chip shipment; the registered sources also conflict on September versus October volume production.
Audit source · page 2 lines 176-181
Microprocessor Report says OEM sampling began in August and volume production was planned for October; it does not establish shipment.
Follow-up: Preserve sampling, planned production, and shipment as separate event types; resolve the timing conflict before inclusion.
Several frozen rows rely on retrospective chronology or contemporary technical reporting without a confirmed first-commercial-shipment record; evidence is insufficient for automatic promotion.
Audit source · chronology lines 4-14
The evidence establishes launch/announcement wording but leaves exact shipment, channel, and regional scope unresolved across multiple rows.
Follow-up: Require human review and targeted primary or contemporary channel corroboration; do not normalize by inference.
These notes preserve the state at intake. Later verification is shown under Release evidence above; an original note saying verification had not occurred is not the current assessment.
PC Watch / Impress ·
Imported source reference; independent verification has not occurred.
Microprocessor Report ·
Imported source reference; independent verification has not occurred.
Chip identity, not the RAGE MAGNUM or RAGE FURY board. Contemporary reporting says ATI announced the RAGE 128 family on August 27 and distinguishes the 128-bit GL chip; sampling is reported, with volume production planned rather than established.